221 13 13MB
English Pages 380 [523] Year 2011
Five years – and one global financial crisis – on from the publication of the highly acclaimed first edition, industry expert Hai Xin has fully revised and updated Currency Overlay drawing on the lessons and experience of the financial crisis. Written by a practitioner for practitioners, the book surveys the major changes in the currency-overlay industry and expands on the practical choices and solutions employed by investors and currency-overlay managers in dealing with real-life problems.
Currency Overlay A Practical Guide Second Edition by Hai Xin
By Hai Xin
New sections provide: • Up-to-date assessments of carry and trendfollowing strategies; • Practical choices for passive overlay, active overlay and currency pure alpha; • New approaches to emerging market currencies; • Reasons behind the increasing use of currency beta indexes; • Detailed discussion on the concept and practice of alpha-beta separation; and • New overviews of markets, participants and legal documents
Hai Xin’s Currency Overlay: A Practical Guide provides a guide to currency-overlay management for all investment portfolios – from a practitioner’s viewpoint. It is the reference work for investment managers whose investment decisions and results are affected by currency fluctuations and who use, or plan to use, an overlay method to manage currency risks.
Currency Overlay A Practical Guide Second Edition
Hai Xin’s Currency Overlay: A Practical Guide shows the reader how to establish a framework to manage currency risk and how leading research into currency trading fits into this framework.
PEFC Certified This book has been produced entirely from sustainable papers that are accredited as PEFC compliant. www.pefc.org
Currency OverlaySBK.indd 1
23/05/2011 14:03
Currency Overlay Second Edition
CO2_with_index_2nd_amends.indd 1
23/05/2011 14:06
CO2_with_index_2nd_amends.indd 2
23/05/2011 14:06
Currency Overlay: A Practical Guide Second Edition
by Hai Xin
CO2_with_index_2nd_amends.indd 3
23/05/2011 14:06
Published by Risk Books, a Division of Incisive Financial Publishing Ltd Haymarket House 28–29 Haymarket London SW1Y 4RX Tel: + 44 (0)207 484 9700 Fax: + 44 (0)207 484 9797 E-mail: [email protected] Sites: www.riskbooks.com www.incisivemedia.com © 2011 Incisive Media Investments Ltd. ISBN 978 1 906348 54 0 British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library
Publisher: Nick Carver Comissioning Editor: Lucie Carter & Jade Mitchell Managing Editor: Lewis O’Sullivan Designer: Lisa Ling Typeset by Tricolour Design, Sudbury, Suffolk Printed and bound in the UK by PrintonDemand-Worldwide Conditions of sale All rights reserved. No part of this publication may be reproduced in any material form whether by photocopying or storing in any medium by electronic means whether or not transiently or incidentally to some other use for this publication without the prior written consent of the copyright owner except in accordance with the provisions of the Copyright, Designs and Patents Act 1988 or under the terms of a licence issued by the Copyright Licensing Agency Limited of 90,Tottenham Court Road, London W1P 0LP. Warning: the doing of any unauthorised act in relation to this work may result in both civil and criminal liability. Every effort has been made to ensure the accuracy of the text at the time of publication, this includes efforts to contact each author to ensure the accuracy of their details at publication is correct. However, no responsibility for loss occasioned to any person acting or refraining from acting as a result of the material contained in this publication will be accepted by the copyright owner, the editor, the authors or Incisive Media. Many of the product names contained in this publication are registered trade marks, and Risk Books has made every effort to print them with the capitalisation and punctuation used by the trademark owner. For reasons of textual clarity, it is not our house style to use symbols such as TM, ®, etc. However, the absence of such symbols should not be taken to indicate absence of trademark protection; anyone wishing to use product names in the public domain should first clear such use with the product owner. While best efforts have been intended for the preparation of this book, neither the publisher, the author nor any of the potentially implicitly affiliated organisations accept responsibility for any errors, mistakes and or omissions it may provide or for any losses howsoever arising from or in reliance upon its information, meanings and interpretations by any parties.
CO2_with_index_2nd_amends.indd 4
23/05/2011 14:06
Contents
List of Figures
xv
List of Tables
xix
Foreword
xxi
About the Author
xxiii
Preface to Second Edition
xxv
Acknowledgements
xxxi
1 1.1
1.2
1.2.1
1.2.2
1.2.3
1.2.4
1.3
1.3.1
1.3.2
1.3.3
1.3.4
1.3.5
1.3.6
1.3.7
1.4
1.4.1
1.4.2
1.4.3
1.4.4
Introduction to Currency Risk Management and Overlay A brief history of modern currencies
1 2
Some stylised facts about the foreign-exchange market
5
Types of trade, transparency and transaction cost
6
Who uses it and why
Market size, major currencies and concentration
5
8
Recent trends in OTC and exchange-traded currency derivatives
12
Absolute return when underlying investment return is certain
15
Absolute return when underlying investment return is uncertain
19
Ex ante versus ex post return
23
Unhedged return, hedged return and local-currency return
25
Short-term versus long-term contribution
26
Stock prices and currency fluctuations
36
Defining currency returns
Selecting a suitable return measure
Relative return
Nominal return versus real return
A simple descriptive tour of currency risks
Correlation
Bond prices and currency fluctuations
13
18
22
24
26
35
38
v
CO2_with_index_2nd_amends.indd 5
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
1.4.5
1.5
1.5.1
1.5.2
1.5.3
1.5.4 1.5.5
1.5.6
1.5.7
1.5.8
1.5.9
1.5.10
1.6
1.6.1
1.6.2
1.6.3
1.6.4
1.6.5
A word of caution for unsuspecting readers of statistics
40
“Currency movements represent random walks and are unpredictable”
42
“Currency is part of global diversification, but hedging reduces diversification benefit”
45
“Currency trading is a zero-sum game”
“Foreign exchange is not a separate asset class”
“Unlike equity or bonds, currency does not attract a risk premium”
2.3
2.3.1
2.3.2
2.3.2.1
2.3.2.2
2.3.2.3
2.3.3
2.4
2.4.1
2.4.2
2.4.3
2.4.4
2.4.5
46
Autocorrelation (momentum, trends and countertrends)
50
Summary
51
Continued interest in foreign-exchange risk management
52
Should currency risk be hedged? Some questions and answers
56
Currency overlay: what it is and what it is not
69
Forward bias (uncovered interest-rate parity, or carry trade)
Purchasing-power parity (PPP) prevails in the long run
FX risk management and currency overlay
The growing global market and its consequences for investors
Survey of studies on currency hedging
Appendix 1C: Forecast, forward and no-arbitrage price
2.2
43
47
Appendix 1B: The correlation triangle
2
41
“Currency market is efficient”
Appendix 1A: Uncovered interest rate arbitrage
2.1
39
What is special about currencies?
48
50
52
55
59
70 71 72
Currency-Overlay Management
77
Origins and development of currency overlay
78
A do-nothing approach
81
A passive hedging approach
83
A dynamic hedging approach
85
Definitions and merits of currency overlay
86
The difference between currency overlay and currency hedging
90
Risk separation and risk responsibility
93
Introduction
Different approaches to managing currency risks
Active versus passive versus dynamic hedging
An active hedging approach
Currency-pure-alpha approach
A working definition
Different types of currency overlay
Competency-based approach
77
81
82
84
86
88
91
94
vi
CO2_with_index_2nd_amends.indd 6
23/05/2011 14:06
CONTENTS
2.4.6
2.4.6.1
2.4.6.2
2.5
2.5.1
2.5.2
2.5.3
2.5.4
2.5.5
2.6
2.6.1
2.6.2
2.6.3
Advantages and disadvantages of employing currency overlay
96
Disadvantages
97
The process of defining a currency overlay
98
Advantages
What an overlay process looks like in practice
A currency overlay laid upon different asset classes
Decision making within an active currency-overlay manager
104
Fundamental-quantitative: the failure of macroeconomics
106
Fundamental-discretionary: on the wane but back with a vengeance
108
A daily flowchart for a currency-overlay manager
Trading styles for active overlay and currency pure alpha
Technical-discretionary: gradually replaced by machines
2.7
Overview on the main parties involved in the currency-overlay business
2.7.1.2
2.7.1.3
2.7.1.4
2.7.1.5
2.7.1.6
2.7.1.7
2.7.1.8
2.7.2
2.7.3
2.7.4
2.7.5
2.8
2.8.1
2.8.2
2.8.3
2.8.4
2.8.5
2.9
2.9.1
2.9.1.1
2.9.1.2
98
101
Technical-quantitative: gaining in importance but suffered badly during crisis
2.7.1
97
The information flow between different parties
2.6.4
2.7.1.1
96
Currency-overlay managers
102
104
107
108 109 109
Internal and external currency-overlay managers
109
Major currency-overlay managers
112
Return-enhancement versus risk-control objectives
114
Trading styles
117
End-users of currency overlay
120
Custodians
123
Historical performance of currency-overlay programmes
124
Two earlier studies of manager performance in 2000
127
Problems with performance-based benchmarks
135
Outsourcing currency management
140
Investment management agreement
141
Overview
Services provided by currency-overlay managers
AUM and AUME figures
Recent changes
Counterparty banks and prime brokers
Consultants
Foreign-exchange-only performance databases and indexes
An analysis of the Parker FX Index
New evidence: 2003–2010
Legal documents for an outsourced currency-overlay programme
ISDA and exceptions
110
113
115
118
121
124
126
128
136
141
143
vii
CO2_with_index_2nd_amends.indd 7
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
2.9.1.3
2.9.2
Prime brokerage agreement
144
Cost of currency overlay
144
Transaction costs
146
Forward premium/discount
148
Track record
149
Research
149
2.10
Summary
150
3
Currency Overlay: Some Practical Choices
155
2.9.2.1
2.9.2.2
2.9.2.3
2.9.2.4
2.9.3
2.9.3.1
2.9.3.2
2.9.3.3
2.9.3.4
Management fees and performance fees
Opportunity cost
Evaluating the service
The organisation
Processes
Appendix 2A: Main players in the currency-overlay market
3.1
3.2
3.2.1
3.2.2
3.2.3 3.2.4
3.2.5
3.2.6
3.2.6.1
3.2.6.2
3.2.6.3
3.2.6.4
3.2.6.5
3.2.6.6
3.2.7
3.2.7.1
3.2.7.2
3.2.7.3
3.2.7.4
3.2.7.5
3.2.7.6
3.2.7.7
3.2.8
144
147
149
149
150
151
Introduction
155
Active or passive overlay approach
156
Running a passive overlay combined with a curreny-pure-alpha programme
159
Choices for overall currency management policy
Active currency overlay or curreny-pure-alpha programme
Internal or external currency-overlay programme
156
158
160
Using several managers
160
The use of derivatives
162
Foreign exchange swaps
163
Use of currency options
165
Risk constraints
166
Volatility limit and tracking-error limit
168
Value-at-risk limit
169
Tolerance range around hedge ratio
170
Portion of portfolio allocated to foreign assets
172
Choice of instruments
Forwards versus futures
Non-deliverable forwards
Other types of innovative instruments
“No net short sale”, “no leverage” and other position limits
Leverage limit
Stop-loss limit and maximum drawdown limit
Different restrictions for different types of currency programmes
161
162
164
166
166
168
170
171
viii
CO2_with_index_2nd_amends.indd 8
23/05/2011 14:06
CONTENTS
3.3
Choices for passive overlay
172
Same or different benchmark hedge ratio for bonds and equities
175
Hedging actual exposure
176
Hedging to benchmark-neutral
177
Instrument tenor
178
Hedging cash account
183
Cross-hedging
186
Currency-fixing arrangements and WM/Reuters rates
188
Choices for active overlay
190
Symmetry of constraints
190
Choices for pure alpha
193
Using listed vehicles
194
Target information ratios and target returns
195
3.6
Overlay and prime brokerage
196
4
FX Research and Choice of Currency Management Styles
201
4.1
Introduction
201
Law of one price
203
Problems with absolute PPP theory
205
Empirical testing
205
Other types of fair-value models
208
3.3.1
3.3.2
3.3.3
3.3.3.1
3.3.3.2
3.3.3.3
3.3.4
3.3.5
3.3.6
3.3.7
3.3.8
3.3.9
3.3.10
3.3.11
3.3.12
3.4
3.4.1
3.4.2
3.4.3
3.5
3.5.1
3.5.2
3.5.3
3.5.4
3.5.5
4.2
4.2.1
4.2.2
4.2.3
4.2.4
4.2.5
4.2.6
4.2.7
Benchmark hedge ratio
Hedging benchmark weights or actual exposures?
Hedging benchmark weighting
Hedging horizon
Rebalancing frequency and tolerance band
Rollover date
Proxy hedging
Reporting
Benchmark hedge ratio
Reporting
Funded or unfunded
Setting a risk budget
Reporting
Models based on purchasing-power parity
Absolute PPP
Relative PPP
Examples of application
172
175
176
177
182
186
187
189
190
192
193
195
195
203
204
205
207
ix
CO2_with_index_2nd_amends.indd 9
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
4.3 4.3.1
4.3.2
4.3.3
4.3.4
4.4
4.5
4.5.1
4.5.1.1
4.5.1.2
4.5.1.3
4.5.1.4
4.5.2
4.5.2.1
4.5.2.2
4.5.2.3
4.5.2.4
4.5.2.5
4.5.2.6
4.6
4.6.1
4.6.2
4.6.3
4.6.4
4.6.5
4.6.6
4.6.7
4.6.7.1
4.6.7.2
4.6.7.3
4.6.7.4
4.6.8
4.7
4.7.1
4.7.2
4.7.3
4.7.4
4.8
4.8.1
Back to fundamentals: macroeconomic models of foreign-exchange determination
210
A brief history of fundamental models
210
Commodity currencies
214
Technical analysis
216
Forward bias revisited
218
Effectiveness
220
Why forward bias might exist
223
Filter rules
225
Effectiveness
226
Countertrending and mean-reversion
230
Other types of quantitative forecast model
231
Markov-switching models
233
Principal component analysis, classification and tree methods
236
Flow models
239
Genetic programming
240
Nearest-neighbour methods
243
Some idiosyncratic currency modelling
246
Definition and varieties
247
A few examples and performance
249
Putting research into action
251
Empirical evidence on the validity of fundamental models
The use of fundamental models in practice and new angles of research
Two warhorses: carry and trend
The long history of carry trade
Some improvements to the naïve carry trade
Variations on the theme of momentum
Moving-average rules
Stop-loss and take-profit
Why momentum may exist
Quantitative foreign-exchange modelling in general
Bayesian updating
Market microstructure and high-frequency data
Other data-mining techniques
Seasonality
Neural-network algorithms
Currency beta indexes
Impact of currency beta indexes
Potential problems with beta indexes
Toolkit
213
214
218
218
222
224
225
229
231
231
235
236
239
242
244
247
248
250
251
x
CO2_with_index_2nd_amends.indd 10
23/05/2011 14:06
CONTENTS
4.8.2
4.8.3
4.8.4
4.8.4.1
4.8.4.2
4.8.4.3
4.8.4.4
4.9
Recent performances and market trend
252
Some other practicalities
255
Objectives of forecasting in the financial market
255
Be wary of career forecasters
257
Are currency models dead?
Pitfalls of data-mining
Interpreting the results with care
Summary: relevance of forecast models in a portfolio-management context
Appendix 4A: Central bank intervention
Appendix 4B: Foreign-exchange and cross-border merger-and-acquisition announcements
5
5.1
5.2
5.2.1
5.2.2
5.2.2.1
5.2.2.2
5.2.2.3
5.2.2.4
5.2.3
5.3
5.4
5.5
5.5.1
5.5.2
5.5.3
5.5.4
5.5.5
5.5.6
5.5.7
5.5.8
5.6
5.7
5.7.1
5.7.2
5.8
254
255
256
257 258 259
Optimisation and Hedge Ratio
267
CAPM and currencies
268
Basic ICAPM models
270
Solnik-Sercu currency-hedging model
271
Inflation model
273
CAPM, foreign-exchange hedging and overlay
274
Hedge ratios
280
Dynamic hedge ratios
282
Minimum-variance hedge ratio
284
Parameter uncertainty and structural models
287
Other factors affecting choice of hedge ratio
290
Currency basket hedging
295
From basket into another basket
297
Introduction
Background
Single-factor model
Currency risk premium in the Solnik–Sercu model
Empirical evidence on ICAPM
Global optimisation and the role of currency
Unitary hedge ratio
Black’s universal hedge ratio
Optimisation and hedge ratios: the empirical evidence
Instability of variance–covariance matrixes
Asset–liability optimisation
From basket into a base currency
Sector-based investment and currency management
267
268
270
272
274
276
281
283
285
289
293
295
298
xi
CO2_with_index_2nd_amends.indd 11
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
6
6.1
6.2
6.3
6.4
6.5
6.6
6.7
6.7.1
6.7.1.1
6.7.1.2
6.7.1.3
6.7.2
6.7.2.1
6.7.2.2
6.7.2.3
6.7.2.4
6.8
6.8.1
6.8.2
6.8.3
6.9
7
7.1
7.2
7.2.1
7.2.2
7.2.3
7.2.4
7.3
7.3.1
7.3.2
7.3.3
7.3.4
Benchmark Design and Performance Measurement
301
Introduction
301
Customised benchmarks: risk separation and risk responsibility
304
The currency-surprise approach
307
Holding-based benchmarks
315
Unhedged
316
Partially hedged
320
Momentum benchmark
320
Fundamental benchmark
322
Adjustment style
324
Hedging benchmark weighting
324
Other practical considerations
328
Setting a good benchmark
Dissecting currency returns
The excess-return approach
Passive benchmarks
Fully hedged (unitary)
Dynamic benchmarks
Carry benchmark
Dynamic hedging benchmark
Hedging actual exposure
Hedging to benchmark neutral
Overlay and Alpha Beta Separation
302
306
312
316
319
320
321
322
324
328
331
Introduction
331
Definitions of currency alpha and beta
331
Historical background
333
The rationale behind passive overlay plus pure alpha
336
Comparing active with passive overlay
338
Summary of active overlay versus passive overlay plus pure alpha
341
Alpha and beta for currencies
Beta dressed up as alpha and alpha transformed to beta
Industry trend
Starting with a passive overlay
Adding pure alpha
7.3.5
Key benefits of using a passive-overlay programme plus a currencypure-alpha programme
7.4
The best implementation methods and other considerations
331
332
335
336
340
342 342
xii
CO2_with_index_2nd_amends.indd 12
23/05/2011 14:06
CONTENTS
8
8.1
8.2
8.2.1
8.2.2
8.2.3
8.2.4
8.2.5
8.2.6
8.2.7
8.2.8
8.3
8.4
8.4.1
8.4.2
8.4.3
8.4.4
8.4.4.1
8.4.4.2
8.4.4.3
8.5
9
9.1
9.2
9.2.1
9.2.2
9.2.3
9.2.4
9.2.5
9.3
9.3.1
9.3.2
9.3.3
9.4
9.4.1
9.4.2
9.4.3
9.4.4
Options, Dynamic Hedging and Overlay
345
Option-pricing theory and dynamic hedging
346
A few intuitive takeaway points from the option-pricing theory
348
Why do people use a dynamic hedging approach to currency overlay
350
Using dynamic hedging to protect a “floor”
354
Use of options for currency-overlay programmes
358
Trading opportunities in the options market
361
Does implied volatility systematically overpredict actual volatility?
362
Some volatility-based trading schemes
366
Selling wings volatility
368
Speaking Greek – some closing comments
369
Introduction
Option price is determined through dynamic hedging
Transaction costs and option replication
Dynamic hedging and trend following
The redundancy of options
Skews and smiles
Overview
Volatility and correlation forecasting
Gamma scalping
Trading spot using option skew as an indicator
Emerging-Market Currencies
345
346
349
353
355
359
361
365
367
368
373
Background
373
Special characteristics of emerging-market currencies
374
Should you hedge emerging currency exposures?
376
Hedging issues for investors based in emerging-market currencies
380
Carry in EM currencies
381
Value-based trading in EM currencies
385
Overview
385
Buying call options on emerging-market currencies to capture carry
386
The hedging debate in EM currencies
More recent interest in EM currencies
Hedging selectively
Carry, trend and value in EM currencies
Trend-following in EM currencies
Use options to trade EM currencies
Selling volatility
Using options to forecast spot
374
375
379
381
383
385
386
386
xiii
CO2_with_index_2nd_amends.indd 13
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
9.5
9.5.1
9.5.2
9.5.3
9.6
10
10.1
10.2
10.2.1
10.2.2
10.2.3
10.3
10.4
10.4.1
10.4.2
10.4.3
10.4.4
10.4.5
10.4.6
10.4.7
11
11.1
11.2
11.3
Other related issues
386
Cross-correlation
386
Indexes of risk and risk appetite
388
Liquidity, Counterparty Risk and E Commerce
393
Foreign-exchange liquidity: theory and reality
394
Electronic dealing and its impact on liquidity
396
Counterparty credit risk
401
General observations
403
Continuous linked settlement
404
Single-bank platforms
408
Custodian banks and eFX
411
Basket peg
Summary
Introduction
Theory of foreign-exchange liquidity
Other factors impacting currency liquidity 2000–2010
E-commerce
Straight-through processing
Multibank platforms
eFX and currency management in practice
Conclusions
387
390
393
394
398
403
404
405
410
413
Summary of the main arguments
413
A currency overlay checklist
420
A practical framework of setting benchmark hedge ratios
416
List of currencies
425
Bibliography
435
Index
477
xiv
CO2_with_index_2nd_amends.indd 14
23/05/2011 14:06
List of Figures
Chapter 1 Figure 1.1
Changing structure of foreign-exchange market
Figure 1.3
Hedged and unhedged MSCI EAFE total-return indexes in US dollar and Japanese yen
27
Figure 1.4
Two-year rolling volatilities computed from the MSCI EAFE returns series from Figure 1.3
28
Figure 1.5
Unhedged and hedged JP Morgan global government bond index in US dollar and Japanese yen
30
Figure 1.6
Two-year rolling volatilities computed from JP Morgan government bond return series from Figure 1.5
31
Figure 1.2
Daily turnover of the foreign-exchange market 1989–2010
7 9
Figure 1.7(a) Risk decomposition on foreign investment portfolios: global equity portfolios
33
Figure 1.7(b) Risk decomposition on foreign investment portfolios: global bond portfolios
33
Figure 1.B.1
A geometric illustration of a correlation triangle
71
Figure 2.1
Payoff profile of a risk-management-motivated currencyoverlay programme
88
Figure 2.2
Currency decisions in the investment decision-making process
99
Figure 2.3
A currency-overlay programme over different asset classes
Figure 2.5
The main parties in a currency-overlay programme and the information flow between them
102
Figure 2.6
Operation of the overlay process when there is a separate currency-overlay manager (internal or external)
103
Figure 2.7
Example of an investment process for an active currencyoverlay manager
105
Figure 2.8
Classification of trading styles
106
Chapter 2
Figure 2.4
A currency-overlay programme where the custodian aggregates all currency positions
100 101
xv
CO2_with_index_2nd_amends.indd 15
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
Figure 2.9
Size of the outsourced currency-overlay industry in terms of AUM (size of the circle corresponding to size of AUM)
111
Figure 2.10
Services provided by a currency-overlay manager
113
Figure 2.12
Total excess return of Parker indexes 1985–2010
130
Figure 2.11
Figure 2.13
Overlay managers and their trading styles
Three-year rolling-average excess returns calculated for Parker Indexes 1985–2010
118
132
Figure 2.14
Annualised three-year rolling volatility calculated for Parker indexes 1985–2010
132
Figure 2.15
Three-year rolling information ratio calculated for Parker Indexes 1985–2010
133
Figure 3.1
Bid–offer spreads of currency swaps and benchmark portfolio against the US dollar for different horizons
180
Figure 3.2
Annualised bid–offer spreads of currency swaps and benchmark portfolio against the US dollar for different horizons
181
Foreign-exchange prime brokerage
197
Chapter 3
Figure 3.3 Figure 3.4
Chapter 4
Implementation of currency overlay in the traditional framework and using a foreign-exchange prime broker
198
Figure 4.1
Exchange-rate change versus inflation differentials over different horizon
206
Figure 4.2
An example of a PPP-based trading scheme
208
Figure 4.4
Stylised representation of new fundamental models of exchange-rate determination (factors shown are not exhaustive)
212
Figure 4.3
Historical performance of value strategies 2000–2010
209
Figure 4.5
Screen shot of UBS’s technical-analysis website
219
Figure 4.7
Historical performance of momentum strategies 2000–2010
227
Figure 4.6 Figure 4.8 Figure 4.9 Figure 4.10
Historical performance of carry strategies 2000–2010 Illustration of a neural-network model
Performance of Parker Systematic FX Strategy Index versus Parker Systematic FX Strategy Index 2003–2010 Deutsche Bank beta indexes during global financial crisis
221 245 252 253
xvi
CO2_with_index_2nd_amends.indd 16
23/05/2011 14:06
LIST OF FIGURES
Figure 4.11
Scattergrams of daily returns calculated from Deutsche Bank G10 Carry Index, MSCI Emerging Market Equity Index March 2003 to September 2010
254
Figure 5.1
Full portfolio optimisation
276
Figure 5.3
Separate currency optimisation
279
Chapter 5 Figure 5.2 Figure 5.4 Figure 5.5 Figure 5.6
Chapter 6 Figure 6.1
Partial portfolio optimisation
Average one-year rolling pairwise correlations for different base currencies, 1980-2010 Asset-only optimisation versus asset–liability optimisation “The virtual world currency” – BNP/OAM Wealth Preservation Currency Index, composition and historical performance
277
291 294
298
Illustration of risk separation
305
Figure 7.1
US dollar evolution against a basket of currencies
334
Figure 7.3
Total portfolio volatility for different passive hedge ratios and for different base currencies
337
Chapter 7 Figure 7.2
Figure 7.4
Figure 7.5
Chapter 8 Figure 8.1 Figure 8.2 Figure 8.3
The evolution of the currency risk-management approach
335
Volatility spread between trend/carry-based active-overlay programmes and passive-overlay programmes for different base currencies
339
Volatility spread between active overlay, passive plus pure alpha and passive-overlay programmes for different base currencies
340
Binomial representation of option-pricing theory
347
Illustration showing how volatilities are linked to option premiums Auto-correlation and estimation of actual volatility
360 364
xvii
CO2_with_index_2nd_amends.indd 17
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
Chapter 9 Figure 9.1
The currency and yield components of EM local currency debt returns
375
Figure 9.2
Carry trading including and excluding emerging-market currencies
383
Figure 9.3
UBS’s FX Risk index + 1998–2010
389
Figure 10.1
Bid-offer spread of 3-month forward contracts in four currency pairs 2007–2010
398
Figure 10.2
Screenshots of the FXall trading platform
407
Figure 11.1
Using optimisation together with practical constraints to set benchmark hedge ratio
417
Figure 11.2
Reactive and proactive approaches to setting a benchmark hedge ratio
418
Figure 11.3
Illustration of how long-term (strategic) and short-term (tactic) currency decisions can be made with reference to each other
419
Chapter 10
Figure 10.3
Chapter 11
Screenshots of UBS’s online eFX tool suites
409
xviii
CO2_with_index_2nd_amends.indd 18
23/05/2011 14:06
List of Tables
Chapter 1 Table 1.1
Currency distribution of global foreign-exchange market turnover – percentage shares of average daily turnover in April 2010 10
Table 1.2
Turnover of OCT and ETD derivatives (US$ billion)
13
Table 1.4
Should currency risk be hedged? A review of the empirical evidence
60
Table 1.3
Table 1.5
UBS Global Asset Management’s investment categories
45
Currency overlay: what it is and what it is not
70
Table 2.1
Historical performance of Parker Global Strategies FX indexes as at October 2010 and February 2003
131
Table 2.2
Correlation between foreign-exchange index returns and underlying asset returns (1990–2000 and 2000–2010)
134
Table 2.3
Recent evidence on whether active currency overlay programmes added value for investors
137
Table 2A.1
Main players in the external currency-overlay services market
151
Chapter 2
Chapter 3 Table 3.1
Restrictions for different currency programmes and typical parameter values 171
Table 3.2
Estimating the cash requirement for forward rolling, assuming a benchmark hedge ratio of 100%
184
Table 3.3
An example of proxy hedging schedule
187
Table 3.4
Chapter 6 Table 6.1
Symmetrical mandate versus asymmetrical mandate
Separating currency returns in the currency-surprise approach
192
307
xix
CO2_with_index_2nd_amends.indd 19
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
Table 6.2
Parameter values for currency-surprise examples
308
Table 6.4
Optimal portfolios using the currency-surprise approach
309
Table 6.3 Table 6.5(a)
Return decomposition using the currency-surprise approach Currency-surprise attribution example: benchmark and actual portfolio allocation assumptions
308
310
Table 6.5(b)
Currency-surprise attribution example: return assumptions
310
Table 6.7
Optimal portfolios using the excess-return approach
313
Table 6.6 Table 6.8
Return attribution in currency-surprise example
Excess returns and cash returns calculated for excess-return approach from data in Table 6.5(b)
312
314
Table 6.9
Return attribution using the excess-return approach
Table 6.11
Calculated raw returns for holding-based benchmark examples
318
Table 6.12
Passive holding-based benchmarks: unhedged benchmark
319
Table 6.10
Table 6.13
Raw data used in holding-based benchmark calculation examples
Passive holding-based benchmarks: fully hedged (unitary) benchmark
315 317
319
Table 6.14
Passive holding-based benchmarks: partially (50%) hedged benchmark
320
Table 6.15
Dynamic holding-based benchmarks: momentum benchmark
321
Table 6.16
Dynamic holding-based benchmarks: carry benchmark
321
Table 6.18
Dynamic holding-based benchmarks: dynamic hedging benchmark
324
Table 6.19
Overlay adjustment: hedge actual exposure
325
Table 6.17
Table 6.20
Dynamic holding-based benchmarks: fundamental benchmark
Overlay adjustment: hedge benchmark exposure
323
326
List of currencies Table A.1
Currencies for which cash/derivatives products are commonly traded
425
Table A.2
Full list of currencies, including precious metals but excluding currencies of euro member states
427
xx
CO2_with_index_2nd_amends.indd 20
23/05/2011 14:06
Foreword
The world of foreign exchange continues to evolve and expand rapidly, at least in terms of global daily average turnover, reaching a remarkable US$4 trillion per day, according to the Bank for International Settlement’s Triennial Central Bank Survey of Foreign Exchange and Derivatives Market Activity in April 2010, reaffirming it as the world’s largest financial market. The structure of the market is changing, with continued consolidation among market-making banks and stratification of bulge bracket banks versus small niche banks; the composition of the market participants is changing, with increasing importance of hedge funds, high-frequency traders, retail aggregators, alongside the traditional real-money asset managers and currency overlay managers; the price discovery mechanism is changing, with electronic platforms and networks revving up their dominance; the whole transaction food chain is changing, with new technologies and services such as straight through processing (STP), continuous linked settlement (CLS), prime-brokerage and so on widely adopted, adding efficiency to the entire workflow from pre-trade, through at-trade to post-trade, while reducing various risks; and the currency trading composition is also changing, with more and more activities in emerging-market currencies, spilling over to frontier currencies. These are all happening around us while we have experienced one of the worst financial crises in living memory, and, as we stand in 2011, we are debating the long-term, or even medium-term, future of two most important currencies in the world: the US dollar and the euro. Will the US dollar crash? Will the euro survive? What would the different answers to these questions mean for the world of foreign exchange and for a particular portfolio? Uncertainty and changes present a mixed bag of good and bad news to investors. Uncertainty may call for a fresh look at the existing currency risk-management policy: some investors may well opt xxi
CO2_with_index_2nd_amends.indd 21
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
for a more conservative approach while others may change tack and try something more adventurous in order to capture the newly emerged opportunities. The new tools, methods and services on the other hand allow the investors to measure, manage and monitor currency risks more effectively and efficiently. But to set up a sensible currency risk-management policy, many choices would need to be made. Some are easier ones but others may have insidious and long-lasting consequences. It is on how to make these choices that I think Hai Xin’s Currency Overlay: A Practical Guide (Second Edition) will likely be of great help to many investors. Hai Xin worked on the first edition around the millennium while working for me as a financial engineer. At the time, the concept and practice of currency overlay was not widely known, but the topic of how to manage currency risk in a global investment portfolio was receiving more and more attention by institutional investors, as many of them were increasing their allocation to foreign-currency-denominated assets. Hai Xin’s book helped many investors understand the ins and outs of currency overlay. In the second decade of the millennium, the world’s goods market and financial market have become much more globalised, and currency overlay has already firmly established itself as one of the standard tools for institutional investors. After leaving the sell side, Hai Xin has spent six years with a currency-overlay manager on the buy side, and I trust that his experience and insight gained from both the buy and sell sides of the foreign-exchange business will make Currency Overlay: A Practical Guide (Second Edition) a comprehensive, objective and relevant book for education, practical advice and reference. Martin Wiedmann Global Head of Foreign Exchange Sales and Distribution Credit Suisse February 2011
xxii
CO2_with_index_2nd_amends.indd 22
23/05/2011 14:06
About the Author
Hai Xin is based in Hong Kong, a managing director and head of Asia Pacific at Overlay Asset Management (OAM) – one of the BNP Paribas Investment Partners. He joined the firm in 2005, focusing on business development and client servicing relating to currency overlay programmes in the region. Before that, Hai Xin worked for the investment bank of UBS (previously Swissbank Corporation, SBC Warburg, SBC Warburg Dillon Read, Warburg Dillon Read, and UBS Warburg – a succession of names that have become a small footnote in the consolidation of the banking sector) in a variety of positions: financial engineer within the Foreign Exchange division, advising corporations and institutional investors on currency hedging and investment issues; “permanent insider” within the Corporate Finance division, advising on currency risk management in cross-border merger and acquisition transactions and as an originator within Equity Capital Market Group dealing with Hong Kong and overseas initial public offerings (IPOs) as well as equity block trades by Chinese companies. Prior to joining UBS he worked as a researcher at AstraZeneca Treasury and as a relief producer for various current affairs programmes at the BBC World Service. Before leaving for Hong Kong, Hai also opened his own tofu factory based in the UK. Hai obtained his PhD in quantitative finance from Imperial College, London, and an MSc in economics from the London School of Economics, both part of the University of London. His bachelor’s degree is in electronics and information science, from the University of Science and Technology of China. Hai has been an adjunct faculty member of the Department of Finance at the Hong Kong University of Science and Technology since 2010, lecturing on statistics as well as investment banking. Hai has published three finance and investment books aimed at the lay reader in China: one being a general introduction to finance and investment, the other two about quantitative investment and xxiii
CO2_with_index_2nd_amends.indd 23
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
high-frequency trading. Hai is also working with Risk Books on The Renminbi Handbook, which will examine the issues relating to trading, investing and hedging of the Chinese Renminbi against other currencies.
xxiv
CO2_with_index_2nd_amends.indd 24
23/05/2011 14:06
Preface to Second Edition
After seven years you are said to get an itch. Mine was to revisit the subject that I worked on seven years ago, bedridden then, having broken a leg playing volleyball in Cuba. The currency-overlay industry has changed, in response to changes in both investors’ requirements and the foreign-exchange marketplace. Personally, I had switched from a salesperson selling products to currency-overlay management companies to an employee in a currency-overlay management company. A fresh perspective of the industry, together with “insider information” about the currency hedging debate, empirical evidences, practical solutions and market trends, made my fingers itchy. The result is Currency Overlay: A Practical Guide (2nd Edition), pruned in some places and significantly expanded in many others, fully updated. WHAT’S NEW The 2nd Edition surveys the major changes in the currency-overlay industry; expands on the practical choices, solutions and, frequently, fudges made and employed by investors and currency-overlay managers in dealing with real-life problems; and addresses some of the imbalances in the 1st Edition by condensing or purging the superfluous contents relating to market views, directory information and currency options. More specifically, here is a list of what is new in the 2nd Edition: o an updated overview of the foreign-exchange market (Chapter 1, Sections 1.1 and 1.2); o an updated summary and discussion of currency risk in international portfolios and some practical tips on how to interpret such statistics (Chapter 1, Section 1.4); o a fully updated survey of empirical studies on the currency hedging debate (Chapter 1, Section 1.6.4);
xxv
CO2_with_index_2nd_amends.indd 25
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
o a restructured explanation on the use of currency-overlay method and the concept of risk responsibility (Chapter 2, Section 2.4); o a new overview of the main parties involved in currency overlay: end-users, currency-overlay managers, counterparties, custodians and consultants (Chapter 2, Section 2.7); o a fully updated performance review of currency-overlay managers (Chapter 2, Section 2.8); o a new overview on the legal documents needed for currencyoverlay programmes (Chapter 2, Section 2.9.1) o an expanded section on the (hidden, and potentially high) transaction costs in running currency-overlay programmes (Chapter 2, Section 2.9.2.2); o new sections on the pros and cons of using active versus passive, active versus pure alpha, and internal versus external overlay programmes (Chapter 3, Sections 3.2.1 to 3.2.4); o new sections on different risk constraints for different types of currency-overlay programme (Chapter 3, Section 3.2.7); o new sections on practical choices for passive overlay, active overlay and currency pure alpha programmes respectively (Chapter 3, Sections 3.3, 3.4, and 3.5); o an update on two main currency trading strategies: carry and trend-following, including their recent performances (Chapter 4, Section 4.5 and Section 4.8.2); o a new section on currency beta indexes and their impact on the currency-overlay industry (Chapter 4, Section 4.7); o an updated section on foreign currencies in asset–liability optimisation (Chapter 5, Section 5.6); o an updated section on basket hedging, both into a base currency or into another basket (Chapter 5, Section 5.7); o a new chapter discussing the concept and practice of alpha–beta separation in currencies, trying to answer the question whether “active currency overlay is less risky than currency pure alpha” (Chapter 7); o an expanded discussion on the dynamic hedging approach to currency overlay (Chapter 8, Section 8.2); o a fully updated section on the hedging debate in emerging-market currencies: what makes it special (Chapter 9, Section 9.2); o a new survey on the carry, trend and value strategies in emerging-market currencies (Chapter 9, Section 9.3); xxvi
CO2_with_index_2nd_amends.indd 26
23/05/2011 14:06
PREFACE TO SECOND EDITION
o a new section on credit risk management (Chapter 10, Section 10.3); and o a new proposal for a framework of setting passive benchmark hedge ratios in practice (Chapter 11, Section 11.2). ABOUT THIS BOOK This book provides a guide to currency-overlay management for investment portfolios from a practitioner’s viewpoint. Our aim is that it should be a reference work for investment managers whose investment decisions and results are affected by currency fluctuations and who currently use, or plan to use, an overlay method to manage their currency risks. It can also be used by plan sponsors/ trustees/managers of pension funds when they consider whether to apply currency overlay on their portfolios and how to outsource its management. The concept and techniques of currency overlay can also be applied to insurance funds, central banks, sovereign wealth funds, endowment funds, family offices, multinational corporations and funds of hedge funds. Currency trading from the currency-overlay industry accounts for a considerable proportion in daily foreign-exchange turnover; therefore, knowing how the currency-overlay industry works may also be useful for sales, structuring and research functions within the banking industry. The concept and practice of currency overlay is not new. Currency risk is an unavoidable consequence of the rising volume of global investment, which has been one of the main themes in investment management for the last 30 years or so. Currency overlay has been gaining acceptance over the last 20 years as more and more investment managers invest abroad and their performance comes under ever closer scrutiny. This book aims to provide guidance on how we should establish a framework to manage currency risk in practice and how various research results on currency trading fit into this framework. As the book is a practical guide, we assume minimum technical background on the part of readers apart from a basic knowledge of investment management. However, we provide references wherever appropriate for more technical readers. STRUCTURE OF THE BOOK To establish whether currency hedging should be used for a parxxvii
CO2_with_index_2nd_amends.indd 27
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
ticular portfolio and to develop an appropriate currency-overlay programme requires consideration of a number of topics. These are listed below along with the chapter in which each is discussed: o the impact of currency on the portfolio’s risk and return (Chapter 1); o the special features of currency versus other asset classes (Chapters 1 and 4); o whether to adopt a passive hedging policy (Chapter 3); o how to select a benchmark hedge ratio (Chapters 5 and 11); o the suitability of active management (Chapters 2, 3 and 4); o benchmarks and performance measurement (Chapter 6); o whether to use active currency overlay or currency pure alpha (Chapter 7); o emerging-market currencies (Chapter 9); and o relevant online technology and e-commerce (Chapter 10). Chapter 2 is devoted to an overview of currency overlay – what it is, how it is done, who are the main players and so on. Many of the practical decisions an investor or an overlay manager needs to make are included in Chapter 3. Further discussions on many of the points in Chapter 3 are discussed in later chapters. For example, in Chapter 4 we review various foreign-exchange models and modelling techniques and the choice of overlay style. In Chapter 5 we include a theoretical digression on currency’s role in modern portfolio theory as well as in the capital-asset-pricing model (CAPM). We list common approaches to determining an optimal hedge ratio for currencies and discuss their suitability in real life. Chapter 6 deals with designing a currency benchmark and measuring currency-management performance. Chapter 7 addresses the issue of alpha–beta separation and how that is relevant in practice. Chapter 8 focuses on the relevance of option-pricing theory, in particular to dynamic overlay style, with a general overview of the use of derivative instruments in currency risk management. The question of whether to have a separate policy for emerging-market currencies is discussed in Chapter 9. Chapter 10 includes a short excursion on currency liquidity and how it varies with changes in the foreign-exchange marketplace, in particular the rapid emergence and spread xxviii
CO2_with_index_2nd_amends.indd 28
23/05/2011 14:06
PREFACE TO SECOND EDITION
of online dealing platforms. We include two examples to show what kinds of services an institutional asset manager can expect from a bank and a multi-bank platform. The book concludes with a summary of the key arguments and a proposed framework in setting benchmark hedge ratios, followed by a step-by-step checklist for someone who is intending to set up a currency overlay. RELATED PUBLICATIONS Currency Overlay, by a well-respected veteran currency-overlay manager, Neil Record (2003), is the closest to this book in scope. Being an insider, Record offered many useful insights about the industry and how currency-overlay programmes work in practice. A few investment banks have published research papers dealing with the currency hedging issue and currency-overlay method over the recent years, notably Folkerts-Landau (2002), A Guide to Currency Overlay Management, and Normand, de Kock, Franklin-Lyons and Sandilya (2010), Managing FX Hedge Ratios: A framework for strategic and tactical decisions. Another Risk publication, Foreign Exchange: A Practitioner’s Approach to the Market (2009), edited by Amy Middleton, also contains some up-to-date discussions on many issues relating to active currency strategies. This book aims to be more comprehensive in topics covered, more structured around the single subject area of currency-overlay management, more neutral in stance by exploring both sides of arguments, and more practical by explaining the choices that an investor or a currency-overlay manager may face in real life and consequences of their decisions. Four earlier publications, all from the Association for Investment Management and Research (AIMR), cover aspects of the topic similar to those dealt with in this book. Managing Currency Risk (1997b) and Currency Risk in Investment Portfolios (1999) are conference proceedings offering a potpourri of contradicting opinions by an assortment of academics and practitioners, usually overlay managers. Ramaswami’s (1993) monograph Active Currency Management, which looks at the non-linear dependencies in current returns and profitable trading strategies, is far too empirical and data-dependent for the purpose of investigating currency risk management in the context of global investment. The closest in objectives and xxix
CO2_with_index_2nd_amends.indd 29
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
layout to our book is Clarke and Kritzman (1996), Currency Management: Concepts and Practices. A tutorial with reviews on currency-management concepts and principles, it presents an algebraic representation of currency hedging in a mean–variance setting (assuming stable multivariate normal distributions) and tables of risks and returns of historical hedged versus unhedged performance. But many important topics are not covered and the restrictive assumptions employed may limit its practical relevance. Four other books on currency risk management cover a similar range of topics to this book. Eun and Resnick’s (2001) International Financial Management provides a comprehensive, if academic, exposition but only a very short section on currency risk in international investment. The Currency Hedging Debate (Thomas III 1990) contains a collection of 18 articles, some previously published (in 1986 and 1989), by both academics and practitioners. Some of these have been reviewed in Section 1.6.4.
xxx
CO2_with_index_2nd_amends.indd 30
23/05/2011 14:06
Acknowledgements
I am grateful to many former and current colleagues at Overlay Asset Management or its parent company, BNP Paribas Investment Partners, for assistance, encouragements and discussions. In particular, I wish to thank Sebastien Gaudre for painstakingly updating all the old charts as well as making some new ones contained in the book; Helie d’Hautefort for his consistent support for this project in more ways than one; Xavier Lefevre for his views and insights on market liquidity; and Hangping Tong, for pointing out many errors in the manuscripts. I would also like to thank Joris Bastien, Xavier Briant, Andrew Broadhurst, Ying Chen, Veronica Dekrey, Edouard Deugnier, Daryle Ee, Simon Godfrey, David Grybas, Robert Harrison, Eglantine Yuanjing Jing, T.G. Lee, Minghui Liao, Lawrence Lo, Ashid Mahmood, Chantal Mazzacurati, Elizabeth Para, Henrik Pedersen, Marian Poirier, Hugues Rondouin, Sebastien Rougier, Fei Sha, Jehanne Suor, Eric Tsang, Ivan Tsui, Oscar Volder, Toshio Wada, Janice Wu and Max Dongyan Xu. I would also like to thank the following people for their help and useful discussions: Stephanie Aufan, UBS; K.C. Lam, Chicago Mercantile Exchange Group; Robert Garwood, Lloyds Banking Group; Elke Kroll, JP Morgan; Martin Wiedmann, Credit Suisse; Justin Snyder and Virginia Parker, Parker Global; Talia Prais-Geberovich, FXall; Michael Huttman and Amy Middleton, Millennium Global Investments Limited; Yue Wu, Crédit Agricole CIB; Dr Chang Shu, Hong Kong Monetary Authority; Felix Stephens and Mark Vrkic, BT Financial Group; Neil Record, Record Currency Management; Terry Yingxia Deng and Dr Yuhong Yan, Hong Kong Securities and Futures Commission; Felix Adam, ACT Partner. The current and former staff members at Risk Books have been persistent, patient and totally professional in moving this project along the way at a reasonable pace and for correcting some of my “Chinglish” or idiosyncratic expressions. I would like to thank the advice and kind assistance by Sarah Hastings, Jade Mitchell and Lewis O’Sullivan. I would also like to thank Clare Beesley and Lucie Carter, who used to work at Risk Books, for following up diligently with me over a period of over two years during the financial crisis, when pretty much everything was up in the air. Finally, I would like to thank my parents. My mum told me that I’d
xxxi
CO2_with_index_2nd_amends.indd 31
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
better work hard on it because “it is so bloody expensive”. I thank Zhen Wu for all the kind help. Any errors and omissions are entirely my own responsibility. Hai Xin [email protected] March 2011, Hong Kong
xxxii
CO2_with_index_2nd_amends.indd 32
23/05/2011 14:06
1
Introduction to Currency Risk Management and Overlay
In this chapter we give some background information on currencies, the currency market and currency risk, present different definitions on currency returns, then review various issues linked to currency hedging and currency risk management. The chapter starts with a brief summary of the history of modern currencies, from Bretton Woods to the aftermath of the global financial crisis. We then look at some stylised facts about currency markets: the players, liquidity, types of transaction and so on. Section 1.3 contains a review of the basic definitions of return in the currency management context. Some simple examples are used to show that different calculation methods can give different results, where the differences can easily be comparable in magnitude to the expected returns from an active currency-overlay programme. We therefore draw up some guidelines on the most suitable definitions of return for use in currency management. We then proceed in Section 1.4 to a descriptive tour of currency risks and correlations in global portfolios. We limit our attention to results that are not too sensitive to different sample periods or calculation methods – namely, the risk reduction that is achievable through currency hedging – but do not consider historical hedged versus unhedged returns. For one thing, we think that conclusions based on those historical return measures are less reliable; For the other, there are plenty of published materials in the area, and some analytical software and tools allow this type of analysis to be conducted on the fly. Section 1.5 presents a list of 1
CO2_with_index_2nd_amends.indd 1
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
claims or questions related to the so-called “currency hedging debate”. Each is reviewed and its relevance to the debate is evaluated. In later chapters we will return to many of these questions. The issue of currency hedging and currency risk management in the context of global investment is discussed in Section 1.6, where we provide an extensive survey of important studies on the usefulness of currency hedging done by both academics and practitioners. A more detailed discussion of currency overlay, its many definitions and practical aspects is postponed until the next chapter. 1.1 A BRIEF HISTORY OF MODERN CURRENCIES The Bretton Woods Conference in 1944, held just as the Second World War was coming to an end, is one of the most important landmarks in the history of modern currencies. Emerging from the ashes of the old system of free-floating gold- or silver-anchored currencies, the new international monetary system was designed to provide stability for war-ravaged economies. Its key ingredients were the convertibility of the US dollar into gold at a fixed rate and the establishment of the International Monetary Fund (IMF) to monitor the new system. Other countries were obliged to peg their respective currencies to the US dollar. The “par levels” were periodically “re-fixed” (devaluation and revaluation by another name) to reflect the misalignment of values, but the whole arrangement essentially created an otherwise fixed exchange system in which central banks intervened on both sides of the currency market to maintain the peg. The Bretton Woods system is sometimes called “US dollar-based gold-exchange standard”. Trading activities, on a minuscule scale compared with today’s, were usually concentrated around the time when these “re-fixes” occurred. In 1967 the Bank of England defended the peg between the British pound and the US dollar against speculators. Although a few other central banks entered the fray to assist the Bank of England, for the first time since the Bretton Woods Conference the speculators won the battle. Unfortunately (or fortunately, as many a free-marketthinking economist would argue), this was not the last time. A new era had dawned, and it was not long before the battered US dollar– gold convertibility was suspended – in 1971, by President Nixon, after several more attacks on the fixed-rate system by the speculators. 2
CO2_with_index_2nd_amends.indd 2
23/05/2011 14:06
Introduction to Currency Risk Management and Overlay
The new floating-rate system commenced two years later, in 1973, by which time the advance of electronic communications technology, coupled with the large increases in both international trade flows and capital flows, had set the stage for the new global foreign-exchange market. On that stage there was an ever-expanding cast of importers/exporters, speculators, banks, exchanges and institutional investors alongside the central banks. Towards the end of the 1970s currency trading was dominated by large banks and brokers. The end of that decade also saw the emergence of the European Monetary System (EMS), in which the new currency unit, the ECU, was introduced to represent a basket of participating European currencies and the member currencies were managed around ECU parity within a prespecified band. The system was repeatedly attacked and was effectively abandoned in 1992–93, when the British pound and Italian lira made their infamous exits. The first half of 1980s was characterised by the steady appreciation of the US dollar against a host of other currencies, but this trend was reversed in the second half of the decade, expedited by the Plaza Accord in 1985 and ensuring coordinated interventions by the Group of Five (US, Japan, Germany, UK and France). Interventions in the currency markets were seen as necessary by many, and they were regularly carried out by major central banks to induce currencies to move in a particular direction. This type of activity was much less common in the 1990s, when the general view was that central banks should not intervene except perhaps to dampen volatility in the short term, and that such intervention should be sparing and effective and have little impact on the long-term equilibrium value of a currency. In 1992 European countries signed the Maastricht Treaty, which drew up the blueprint for the new single currency as well as the unified monetary policies for the member states. In the ensuing few years leading to the debut of the euro in 1999, member countries struggled to fulfil the criteria for joining the single currency. The euro had had a tumultuous gestation, a reasonably smooth birth and a loveless infancy, and nose-dived from 1.2 to one US dollar at the beginning of 1999 to around 0.80 to one US dollar two years later. Euro began to have what seemed like an existential problem – its first one, but not the last. 3
CO2_with_index_2nd_amends.indd 3
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
The 1990s also witnessed the phenomenal growth in foreign-exchange derivatives, the complete dominance of electronic trading/ broking at the expense of other dealing methods, and the wholesale consolidation of banks active in foreign-exchange trading. Foreign exchange became a mature industry, with a commoditised product range and razor-thin bid–offer spreads. The volume in the foreignexchange market, in terms of both cash and derivatives, began to fall in the last few years of the old millennium, induced probably by the appearance of the euro and a gradual withdrawal of speculative capital (which flew to stock markets around the world). There were fewer banks active in currency market and, in every one of them, fewer employees. Some market insiders were predicting the slow and anaemic death of foreign-exchange business. The talk of a single world currency, toyed with almost a hundred years ago when gold acted like one, resurfaced, but this time people were talking about the ubiquitous “e-currency”, whereby a countryless and borderless electronic point system would assume most, if not all, roles traditionally fulfilled by currencies, like the chips in a casino – this particular casino being the World Wide Web. That proved to be a tulip-shaped bubble (much to my and my colleagues’ relief). Money from investors and speculators was back in the foreignexchange market in 2002 and 2003, with other investment arenas offering lean pickings. The decisions taken by the Federal Reserve to slash US interest rates by 550 basis points between 2001 and 2003 pushed the US dollar on a downward trajectory. Some currency speculators had made oversized gains from their short US dollar positions around 2003, riding on a smooth one-way train. The ensuing years, 2004 to 2007, when China emerged on the central stage of international trade and finance and before the global financial crisis confronted us totally unprepared, heard talks aplenty about the “Great Moderation”, whereby many currency pairs stuck in narrow ranges with periodical realignments and the currency volatility went lower and lower. Speculators in the currency market either piled up on carry trade or sold volatilities, or, indeed, like the fabled Japanese housewife Mrs Watanabe, with a popular product called “dual currency deposit”, did both. Many regretted their actions when the tsunami flattened the fantasy world we all lived in, built on tranche upon tranche of debt, in its myriad acronymic guises. 4
CO2_with_index_2nd_amends.indd 4
23/05/2011 14:06
Introduction to Currency Risk Management and Overlay
During the financial crisis, the interbank market – the backbone for the global currency market – nearly froze, when liquidity disappeared and transaction costs spiked. Credit risk, which was hitherto not a big concern in FX trading, turned out to be a serious issue. Investors deleveraged and unwound the “carry trades” – an ageold currency-trading tactic whereby investors borrow in low-interest-rate currencies and invest the borrowings in high-interest-rate currencies, and hope that the ensuing currency movements will not wipe off the gain in interest rate differentials – and then releveraged and put on fresh carry-trade positions when the market condition improved; but the yield was low almost everywhere. Investment banks made healthy profits from all these activities, driven by a volatile mixture of fear and greed. One beneficiary of the crisis was exchange-traded FX products, which enjoyed a renaissance in a world full of suspicions. Another beneficiary was emerging-market currencies as a whole, excluding, however, the yet-fully-convertible Chinese renminbi. The tidal waves of hot money created by the ultra-loose monetary policies in the developed world in order to shock their economies out of stupor have poured into the emerging markets, their currencies facing upward pressure and the competitiveness of their export industries, downward – unless, of course, they intervene, tax, raise interest rates and introduce capital controls. Many did. The talk of “currency war” makes regular headlines and we have a US dollar scare one day, followed by a euro scare another. Many questioned the hegemony of the US dollar as the world’s reserve currency. Suggestions on alternatives abound, from “Special Drawing Rights” (SDR) created by the International Monetary Fund, to a “super-sovereign reserve currency”, a soupedup version of SDR, proposed by the People’s Bank of China governor Zhou Xiaochuan in 2009, to gold, suggested one year later by Robert Zoellick, president of the World Bank, but none looked particularly creditable. 1.2 SOME STYLISED FACTS ABOUT THE FOREIGN-EXCHANGE MARKET 1.2.1 Who uses it and why Major participants in the foreign-exchange market include:
5
CO2_with_index_2nd_amends.indd 5
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
commercial banks; investment banks; central banks, sovereign wealth funds, supernational banks; corporations; fund managers (including those of mutual funds, pension funds, insurance funds, currency-overlay programmes, etc.); o hedge funds; and o private individuals. o o o o o
Although one of the original purposes of foreign exchange was to facilitate international trade, only a tiny percentage of the global foreignexchange volume is directly linked to trade and service flows. Marketmarkers, investors, speculators, arbitrageurs and hedgers of varying sizes are active in the currency markets. Some view foreign exchange as a distinctive asset class and seek excess return by buying and selling currencies. Others may not treat foreign exchange as an asset class: rather, their need to buy and sell currencies arises from investment decisions to buy and sell securities denominated in foreign currencies and/or to hedge the currency risks. Multinational corporations buy and sell currencies as part of their day-to-day operations. They may also buy and sell currencies to hedge their currency exposures. Many use currency instruments to manage their funding and short-term liquidity. In addition, there are central banks and similar institutions that may trade currencies towards a particular target level to dampen market volatility or to manage a country’s foreign-exchange reserves. Finally, tourists buy local currencies to spend. Many commentators state that the existence of large numbers of participants that are not aiming at maximising profit from trading foreign exchange presents the profit opportunities for those who are. Yet the empirical evidence on this point is lacking. Some banks estimate that less than 25% of the volume traded on the foreign-exchange market is seeking profit, but that is not evidence in itself that an economically meaningful return could be earned by all or majority of the profit-maximisers. One single carnivorous fish in a pond may need a lot of herbivores for support. 1.2.2 Types of trade, transparency and transaction cost Foreign-exchange transactions can be classified into three main types according to their counterparties: 6
CO2_with_index_2nd_amends.indd 6
23/05/2011 14:06
Introduction to Currency Risk Management and Overlay
1. trades between customers and banks; 2. trades directly between banks (interbank); and 3. brokered trades between banks. Given the size of the market, the volume in each category is difficult to measure and it varies with time and different currency pairs. Historically, around a third of the total trades were customer-related and the remaining two-thirds were interbank trades. However, this ratio has been changing in favour of customer-related trades over recent years as banks have been trying to “internalise” more trades in order to capture bid–offer spread rather than hedging all positions in the interbank market. Also, the consolidation and specialisation of the banks participating in the interbank market means that there are fewer large banks active in the interbank market and many smaller banks have become “customers” of the large banks. This point is illustrated in Figure 1.1. Traditionally there are more direct interbank trades than brokered ones, but the trend over the past few years has been towards a growing proportion of brokered trades, particularly electronically brokered. Figure 1.1 Changing structure of foreign-exchange market
Old FX Market Structure Large Banks
Small Banks
New FX Market Structure Customers
7
CO2_with_index_2nd_amends.indd 7
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
Most foreign-exchange transactions are conducted over the counter (OTC) between counterparties, with one, or, more likely, both of them, being market-makers.1 This market microstructure results in a paradoxical characteristic of the foreign-exchange market: although the trading volume is massive, transparency can be low. Information on a given transaction is usually not known beyond the two parties involved. Foreign-exchange prices distributed by information vendors, such as Reuters or Bloomberg, are not necessarily transaction prices. Bid–offer spreads in the foreign-exchange market are very small, usually a couple of basis points. Market-makers (mostly banks) typically make more money by taking directional positions rather than capturing the bid–offer spread – these banks enjoy an information advantage because of the client flows they have. Academic as well as anecdotal evidence suggests that profit is usually made when the market is thin (eg, during the financial crisis, or for emerging-market currencies in general) and/or volatility is high (eg, Japanese yen in 1997–98, the euro in 2002, and the global financial crisis again) and/or an important announcement is made or about to be made (eg, unexpected change of interest rate). 1.2.3 Market size, major currencies and concentration Foreign exchange is the biggest capital market, with an average daily turnover at US$4 trillion according to the latest Triennial Central Bank Survey of Foreign Exchange and Derivatives Market Activities 2010, published by the Bank for International Settlements (BIS) in September 2010. This represents a 20% gain compared with the volume for 2007. However, US$4 trillion is a massive increase compared with the figure of US$1.2 billion in 2001, the only year that the turnover showed a decrease since the survey started in 1989 (see Figure 1.2). If we look at the trades by their counterparties, the percentage of interbank trades in the total turnover accounted for only 39%, a significant fall from the 2001 figure of 59%, mainly due to the decreasing level of activities of proprietary trading of banks as well as banks’ effort to “internalise” customer transactions by warehousing them. At the same time, trades between banks and financial counterparties increased in share from 28% in 2001 to 48% in 2010. This 8
CO2_with_index_2nd_amends.indd 8
23/05/2011 14:06
Introduction to Currency Risk Management and Overlay
Figure 1.2 Daily turnover of the foreign-exchange market 1989–2010
Daily average volume ( US$ billion)
4,000
Spot
Forward
Swaps
Options and other products
Total
3,500 3,000 2,500 2,000 1,500 1,000 500 0
1989
1992
1995
1998
2001
2004
2007
2010
Source: BIS Triennial Central Bank Survey of Foreign Exchange and Derivatives Market Activities (BIS, September 2010). Available at www.bis.org.
group includes central banks, hedge funds, mutual funds and smaller commercial banks. Many smaller banks now become customers of large international banks, obtaining liquidity, streamed prices, even the IT infrastructures from large banks. Their trading volumes are now captures in customer flows rather than the interbank flows. Trading activities in the hedge fund segment have also increased substantially over the period, especially those using high-frequency trading strategies. The US dollar appears in 85% of the trades as one of the legs, a decrease of 5% over the nine-year period, whereas the euro, Japanese yen and sterling account for 39% (38%), 19% (23%) and 13% (13%), respectively (figures in parentheses are 2001 numbers). The euro/US dollar is still the most frequently traded currency pair, capturing 28% (30%) of turnover, followed by US dollar/yen 14% (20%), sterling/US dollar 9% (11%). The emerging-market currencies have increased their share of total trading volume, albeit from a small base. For example, the Korea won, from 0.8% to 1.5%; Indian rupee, 0.2% to 0.9%; Russian rouble, 0.3% to 0.9%; Chinese renminbi, 0.008% to 0.3%. Table 1.1 shows the breakdown of turnover in different currencies. 9
CO2_with_index_2nd_amends.indd 9
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
Table 1.1 Currency distribution of global foreign-exchange market turnover – percentage shares of average daily turnover in April 20101 Currency
1998
2001
2004
2007
2010
US dollar
86.8
89.9
88.0
85.6
84.9
...
37.9
37.4
37.0
39.1
30.5
...
...
...
...
5.0
...
...
...
...
16.8
...
...
...
...
...
0.0
0.0
0.1
...
Japanese yen
21.7
23.5
20.8
17.2
19.0
Pound sterling
11.0
13.0
16.5
14.9
12.9
Australian dollar
3.0
4.3
6.0
6.6
7.6
Swiss franc
7.1
6.0
6.0
6.8
6.4
Canadian dollar
3.5
4.5
4.2
4.3
5.3
Hong Kong dollar3, 4
1.0
2.2
1.8
2.7
2.4
Swedish krona5
0.3
2.5
2.2
2.7
2.2
New Zealand dollar3, 4
0.2
0.6
1.1
1.9
1.6
Korean won3,4
0.2
0.8
1.1
1.2
1.5
Singapore dollar3
1.1
1.1
0.9
1.2
1.4
Norwegian krone3
0.2
1.5
1.4
2.1
1.3
Mexican peso3
0.5
0.8
1.1
1.3
1.3
Indian rupee3,4
0.1
0.2
0.3
0.7
0.9
Russian Rouble3
0.3
0.3
0.6
0.7
0.9
Polish zloty3
0.1
0.5
0.4
0.8
0.8
...
0.0
0.1
0.2
0.7
0.4
0.9
0.7
0.9
0.7
Euro Deutsche mark French franc ECU and other EMS currencies Slovak koruna2
Turkish new lira2 South African rand3, 4
10
CO2_with_index_2nd_amends.indd 10
23/05/2011 14:06
Introduction to Currency Risk Management and Overlay
Brazilian real3, 4
0.2
0.5
0.3
0.4
0.7
Danish krone3
0.3
1.2
0.9
0.8
0.6
New Taiwan dollar3
0.1
0.3
0.4
0.4
0.5
Hungarian forint3
0.0
0.0
0.2
0.3
0.4
Chinese renminbi4
0.0
0.0
0.1
0.5
0.3
Malaysian ringgit2
0.0
0.1
0.1
0.1
0.3
Thai baht3
0.1
0.2
0.2
0.2
0.2
Czech koruna3
0.3
0.2
0.2
0.2
0.2
Philippine peso2
0.0
0.0
0.0
0.1
0.2
Chilean peso2
0.1
0.2
0.1
0.1
0.2
Indonesian rupiah2
0.1
0.0
0.1
0.1
0.2
Israeli new shekel2
...
0.1
0.1
0.2
0.2
Colombian peso2
...
0.0
0.0
0.1
0.1
0.1
0.1
0.0
0.1
0.1
8.9
6.5
6.6
7.6
5.3
200.0
200.0
200.0
200.0
200.0
Saudi Riyal
2
Other currencies All currencies
Because two currencies are involved in each transaction, the sum of the percentage shares of individual currencies totals 200% instead of 100%. Adjusted for local and cross-border inter dealer double-counting (ie, “net–net” basis). 2 Data previous to 2007 covers local home currency trading only. 3 For 1998, the data covers local home currency trading only. 4 Included as main currency from 2010. For more details on the set of new currencies covered by the 2010 survey, see the statistical notes in Section IV of BIS (2010). 5 For 1998, the data covers local home currency trading only. It has been included as main currency from 2007. 1
Source: BIS Triennial Central Bank Survey of Foreign Exchange and Derivatives Market Activities (BIS, September 2010). Available at www.bis.org.
Another important feature of the foreign-exchange market is its market concentration. There are several aspects to this concentration. o The US dollar continues to be the dominant currency on the global foreign-exchange market, appearing in nearly 85% of the transactions as one leg (the total % of all currencies sums up to 11
CO2_with_index_2nd_amends.indd 11
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
200% because there are two currencies in each transaction). This percentage has barely changed over the last five BIS surveys. o London and New York continue to be the dominant centres of global currency trading activities, accounting for 37% and 18% of the total trading volume. These percentages have been relatively stable over the preceding decade, with London crawling up a few percentage points. Tokyo is a distant third, with 6% of global trading activities, edging down from the 10% it used to capture in 1995. o The number of banks and relative sizes of these banks active on the interbank market have been declining, continuing the trend of consolidation. The global financial crisis helped to expedite the process. Within two trading centres, London and New York, the number of banks accounting for 75% of the turnover was also trending down. 1.2.4 Recent trends in OTC and exchange-traded currency derivatives The volume in the OTC foreign-exchange derivatives market grew in tandem with the overall FX market volume growth; outright forwards and options contracts account for 12% and 5% of the total transaction volume in 2010 respectively, similar proportions to those in 2001. The proportion of FX swaps, historically accounting for around 50% of all FX transactions, declined somewhat to around 44%, mainly due to the lacklustre growth in interbank swaps volumes. This slack was more or less taken up by the growth in spot transactions, ramping up from 31% in 2001 to 37% in 2010. The growth in spot transactions was mainly from outside the interbank market. The volume of FX spot between market-makers and the customers went from US$169 billion per day in 2001 to US$972 billion per day in 2010. Table 1.2 summarises the OTC turnover data for both currency and interest-rate derivatives. We include exchange-traded derivatives (ETDs) for comparison. Several additional observations can be made. o The growth in both OTC and ETD interest-rate derivatives has been at a similar pace to that of OTC FX derivatives. The trading volume in the ETD interest derivatives continues to be several times larger than the OTC counterpart.2 12
CO2_with_index_2nd_amends.indd 12
23/05/2011 14:06
Introduction to Currency Risk Management and Overlay
o The growth in the ETD FX contracts has been explosive, going from a negligible US$12 billion per day in 2001 to US$166 billion per day in 2010, due to the emergence of high-frequency trading as well as the concern over credit risks. We will come back to this point later in the book. Table 1.2 Turnover of OCT and ETD derivatives (US$ billion) 1995
1998
2001
2004
2007
2010
Foreign exchange
684
949
846
1,282
2,288
2,447
Forward/swap
643
862
786
1,163
2,076
2,240
41
87
60
119
212
207
Interest rates
151
265
489
1,025
1,686
2,083
Total OTC derivatives
800
1,263
1,387
2,307
3,974
4,530
Total ETD, of which:
1,222
1,373
2,179
4,550
6,179
8,308
17
11
12
26
80
166
1,205
1,381
2,188
4,524
6,099
8,142
FX options
Foreign exchange Interest rates
Source: BIS Triennial Central Bank Survey of Foreign Exchange and Derivatives Market Activities (BIS, September 2010); ETD data from FOW TRADEdata; Futures Industry Association; various futures and options exchanges. Reported monthly data was converted into daily averages of 20.5 days in 1998, 19.5 days in 2001, 20.5 in 2004, 20 in 2007 and 20 in 2010.
In the OTC foreign-exchange derivatives business, the interbank volume accounted for 39% in 2010. In 2001, the percentage was 58%, again a reflection of the increasing activities of institutional fund managers as well as the consolidation of the interbank market. 1.3 DEFINING CURRENCY RETURNS Many attempts to investigate the currency risk in international investments are either flawed or unnecessarily complicated due to 13
CO2_with_index_2nd_amends.indd 13
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
the lack of a consistent approach. Concepts of return are often confused – for example, ex ante versus ex post returns, spot versus forward returns, domestic versus foreign interest rates, bond/equity returns expressed in one currency versus hedged into the currency, the total return on a benchmark bond portfolio versus the shortterm interest rate in the same currency, etc. In this section we present a brief overview of the relevant concepts of return. Methods to decompose returns on foreign investments into currency returns and underlying returns will be discussed in Chapter 6. Return can be either absolute or relative. The latter is the return computed relative to a specific benchmark. Return can also be ex ante or ex post (see Section 1.3.5). Returns for currencies can be more difficult to define than returns on, say, bonds and equities. So we will first look at a stylised example to help us approach the problem of definition gradually. Example 1.1: A US-dollar-based investor invests US$1 million in a Japanese monetary fund. The market prices are: o spot exchange rate is 100 Japanese yen per US dollar (denoted USD/JPY in subsequent sections);3 o three-month forward exchange rate is 98.46 USD/JPY; o three-month Treasury bill yield is 6.50% (in US dollars); and o three-month Japanese government bond (JGB) yield is 0.25% (in Japanese yen).4 To obtain the yen funds needed to purchase units in the Japanese monetary fund, the investor can conduct a spot currency transaction to convert the US$1 million into yen at the current spot rate. He receives ¥100 million and the proceeds are invested. Three months later the spot exchange rate moves from 100 to 110 USD/JPY and the ¥100 million investment has turned into ¥100.0625 million.5 What is the return on currency? What is the relevance of Treasury bill and JGB yields? And would the result be different if the money were invested in an equity or bond fund? Would the return calculation be different if the currency were fully or partially hedged? We will consider these questions in the next subsections. 14
CO2_with_index_2nd_amends.indd 14
23/05/2011 14:06
Introduction to Currency Risk Management and Overlay
1.3.1 Absolute return when underlying investment return is certain Here are a few commonly used methods, or return measures, for calculating the absolute return on currency. Method I: The first is a straightforward calculation of how much the Japanese yen has depreciated during the three-month period on a spot-on-spot basis. This method is often used by journalists and other financial commentators to calculate currency moves, and the results are sometimes referred to as “headline” currency returns. Initial FX Spot
Initial FX Spot
End FX Spot
Annulisation
Method II: Some people may argue that the first method is more relevant for US dollar appreciation than yen depreciation. To calculate the latter, we need to redenominate the exchange rate as US dollars per Japanese yen rather than the conventional Japanese yen per US dollar, as in Method I. Again, during the three-month period, on a spot-on-spot basis: End FX Spot
Initial FX Spot
Initial FX Spot
Annulisation
The difference between Methods I and II can be attributed to the way percentage returns are calculated. To take an extreme example, an increase from 1 to 10 is 900%, but a corresponding decrease from 10 to 1 is only 90%. In practice, the difference is usually not this 15
CO2_with_index_2nd_amends.indd 15
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
large because the change in a given period is small if not negligible. The difference disappears if one uses log-returns. Method III: Similar to Method I, but instead of spot-on-spot we use spot-on-forward and the exchange rates are expressed in Japanese yen per US dollar: Initial FX Forward
End FX Spot
Initial FX Forward Annulisation
We observe that the return in this case, 46.48%, is lower than the return obtained with Method I. This is because the Japanese yen is trading at a premium to the US dollar. In other words, the yen is “expected” to be stronger in the future. Therefore, its depreciation against the US dollar is even more pronounced when set against this “expectation”. However, the forward price is not an “expectation” of the future spot price in a strict sense – unlike forward interest rates calculated from the yield curve, which is the market forecast of future interest rates. The foreign-exchange spot–forward relationship is based on a no-arbitrage relationship. This is an important distinction but not totally relevant to what we are trying to concentrate on here. For further details see Appendix 2C. Method IV: Similar to Method II, but with the three-month forward rate replacing the initial spot rate. End FX Spot
Initial FX Forward
Initial FX Forward
Annulisation
16
CO2_with_index_2nd_amends.indd 16
23/05/2011 14:06
Introduction to Currency Risk Management and Overlay
Method V: From a cashflow viewpoint, the investor has US$1 million to start with. If they liquidate their investment in Japanese yen at the end of Month 3 and convert the proceeds into US dollars at spot, a return can be calculated from the net return as follows: End FX Spot End Investment
Initial Investment
Initial Investment Annulisation
Method VI: Similar to Method V but with a small adjustment as, strictly speaking, the return in Method V is not a return on currency, because it includes the return on the Japanese yen monetary fund, which yields 0.25% per annum. In fact, the same adjustment can be made to the other methods as well. End FX Spot End Investment
Initial Investment
Initial Investment
Risk-free Interest Rate
Annulisation
It is worrying that the return figures in such a simple example can differ by so much, and none of them seems particularly wrong! The discrepancies are particularly alarming in view of the fact that his17
CO2_with_index_2nd_amends.indd 17
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
torical or expected average returns from active currency management amount only to tens of basis points per annum. If used by unsuspecting portfolio optimisers, different calculation methods could radically alter asset allocation decisions. This is one of the reasons that some empirical studies produce conflicting evidence with the same data set. 1.3.2 Selecting a suitable return measure With a confusing array of alternative measures to choose from, it is important to be clear about which we should use. Although we do not think there ought to be a universal measure suitable for different types of investors under different circumstances, here are a few general guidelines: o it should be consistent with the return measures used for other asset classes and should be readily usable in optimisation together with other asset classes if needed; o it should be capable of reflecting the economic impact of currencies on investments; o it should be investable (hedgeable); a portfolio manager can reproduce the return almost exactly (after taking transaction costs into account) if they choose to by following some predefined rules irrespective of actual market moves; o it should not give a portfolio manager incentive or disincentive to hedge or not to hedge currency risk beyond what is prescribed by a currency benchmark; o it should be able to take the use, and hence cost, of capital into account;6 o it should be relatively easy to apply to both benchmark and actual portfolios for return–risk calculations; and o it should reduce the dead-weight return – the portion of return where neither currency manager nor underlying manager is responsible (see Chapter 6, Section 6.3) Whereas the discrepancy caused by different base currencies is easy to deal with (for example, by using log-returns rather than simple returns), the choice between spot-on-spot and spot-on-forward is less obvious. Many people argue that it is dependent on whether investors hedge their currency exposure: if you hedge, then you use 18
CO2_with_index_2nd_amends.indd 18
23/05/2011 14:06
Introduction to Currency Risk Management and Overlay
spot-on-forward returns; if you do not hedge, then you use spoton-spot returns. It seems to make intuitive sense because those are the types of return you will get. However, the use of different return definitions for different hedging choices complicates the analysis greatly and, perhaps, unnecessarily. In this book, unless specified otherwise, we use log-returns on a spot-to-forward basis and we will go through the reasons in Chapter 6, Section 6.4. Cashflow-based methods (Methods V and VI) are widely used for performance evaluation purposes. As portfolios are marked-tomarket during each period, the returns can then be computed from the difference in valuation. One main advantage of using cashflowbased methods is that the computed returns are intrinsically investable and this is a very important advantage in practice. Whenever possible, we think cashflow-based methods should be used in reallife analysis, even though such methods are usually more complicated to conduct and may be dependent on specific circumstances faced by an investor. However, the change in valuation results from change in the underlying investment and change in currency prices. To separate these effects, we can hold other factors constant and allow only one to vary. 1.3.3 Absolute return when underlying investment return is uncertain Since the return calculation is central to many of the topics in this book, we will explore it a little further. In Section 1.3.1, the return on the underlying investment was fixed and known at the beginning of the investment period. The picture can be further complicated by the fact that the return on the underlying is not known at the beginning. To illustrate this we will rework Example 1.1 using equities. Example 1.2. A US-dollar-based investor invests US$1 million in Japanese stocks. The market prices are: o o o o
spot exchange rate is ¥100 per US dollar; three-month forward exchange rate is 98.46 USD/JPY; three-month Treasury bill yield is 6.50% (in US dollars); and three-month JGB yield is 0.25% (in Japanese yen).
The Nikkei Index stands at 10,000 at the beginning of the investment and goes up to 11,000 (adjusted for dividends) after three months. 19
CO2_with_index_2nd_amends.indd 19
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
As in Example 1.1, the investor can conduct a spot currency transaction to convert the US$1 million into Japanese yen at the current spot rate. They receive ¥100 million and invest the proceeds in local stocks. Let us assume that the investor’s return follows the Nikkei exactly. The spot exchange rate again moves from 100 to 110 USD/JPY and the ¥100 million investment turns into ¥110 million. What is the return on currency? The calculation for Methods I through IV remains the same, since the return on the underlying investment does not feature in the calculation. We will rework Method V here as Method V(b). Method V(b): In cashflow terms, the investor has US$1 million to start with. If they liquidate their investment in Japanese yen at the end of Month 3, they are left with ¥110 million. If they then convert the proceeds into US dollars at the spot rate, a return can be calculated from the net return as follows: End FX Spot End Investment
Initial Investment
Initial Investment
Annulisation
This is clearly wrong, as the Japanese yen has depreciated against the US dollar over the period and the currency return should not be zero. It is not difficult to observe that the underlying investment produces a return of 40%, which subsidises the negative currency return. How should we deal with this? We want a return calculation method that produces a return that is investable. From a foreign-exchange hedging viewpoint, the above return is not investable, as one does not know at the outset what the final Japanese yen asset will be. So, two practical questions need to be asked: 20
CO2_with_index_2nd_amends.indd 20
23/05/2011 14:06
Introduction to Currency Risk Management and Overlay
1. How much investment is exposed to exchange-rate movement (the exposure)? 2. Is the hedge static or it is adjusted during the three-month calculation period? The choice for the first question is usually one of the following: o current valuation, denominated in Japanese yen; in our example this is ¥100 million; o current valuation, denominated in US dollars, ie, US$1 million; o future valuation at the forecast growth rate, denominated in Japanese yen (if we assume that the investor expects the Nikkei to grow by 5% over the three months, then the exposed amount is ¥105 million); o future valuation at the forecast growth rate, denominated in US dollars; this is US$1.0664 million (=105 million/98.46); o future valuation at the Japanese yen risk-free growth rate, denominated in Japanese yen; this is ¥100.0625 million; and o future valuation at the US dollar risk-free growth rate, denominated in US dollars; this is US$1.01625 million. Obviously, each will yield slightly different returns if we recalculate the returns using Method V(b). The second question – whether the exposure (and the hedge) is adjusted during the calculation period – is also relevant. For example, if halfway through the three-month period the Nikkei has gone up from 10,000 to 10,500, the initial exposure amount has clearly changed. Will the investor elect to adjust the exposure or leave the hedge static? If they adjust the exposure, the new calculation of the return will have to reflect the fact that halfway through the period the USD/JPY exchange rate will also have changed. Clearly, the adjustment too will affect the final return calculation and so needs to be defined beforehand. In practice the exposure is usually defined as the current markto-market valuation of the foreign investment denominated in foreign currency, in this case, ¥100 million. The issue of rebalancing is an important one and we will revisit it in Chapter 3, Section 3.3.6.
21
CO2_with_index_2nd_amends.indd 21
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
1.3.4 Relative return Relative returns are more commonly used in the fund-management industry, and here calculation of the return on a currency is even more plagued by ambiguity. One commonly used term is “excess return”. For other types of asset, excess return is defined as the absolute return from investment less the return on a risk-free asset. What, then, is the excess return on a currency? As demonstrated earlier, the definition of the absolute return can vary, and so can the definition of the return on a risk-free asset. These are the usually quoted alternatives: o since the expected currency return is zero, we can use one of the absolute returns obtained in the previous section as the excess return; and o a foreign currency can trade at a premium (more expensively) or discount (cheaply) relative to the spot exchange rate7 in the forward market, and some use the forward point (difference between spot and forward) as the expected return on a currency. F'wd points
Relative return on currency = Absolute return + 1.54 x 4 USD/JPY 1.54 = Absolute return + x 4 x 100% 98.46 = Absolute return + 6.25% Also interest rate differential
Annulisation
Some people follow the general rule of using risk-free rate in the calculation of excess return, but, in our example, which currency should we consider, the US dollar or the yen? This is a very important question in currency risk management and its significance goes well beyond obtaining a more “accurate” return measure. We will defer detailed discussion on this point until Chapter 6, where we look into the question of index calculation and performance measurement.
22
CO2_with_index_2nd_amends.indd 22
23/05/2011 14:06
Introduction to Currency Risk Management and Overlay
1.3.5 Ex ante versus ex post return Ex ante (meaning viewed “from before”) return is the forecast or expected return, which can be used in asset allocation, optimisation and portfolio construction. Ex post (“from after”) return is the realised return, and it can be used in performance attribution and risk management. As we will see later in Section 1.6.4, many earlier studies of currency risk management are ex post. Optimisation techniques are applied to ex post returns and variance–covariance matrixes to examine the impact of currency hedging. The practical relevance of this type of result is very limited, because, when, a portfolio manager makes a trading decision, they have historical price data only up to the point at which they make that decision. So they will have to decide whether to use the available historical data or, instead, use some form of forecast. If the historical analysis is conducted incorporating this constraint, the results can be drastically different from those obtained using a complete ex post dataset. More recent studies take this vital practical constraint into consideration and their results – for instance, on currency optimisation and the use of a unitary hedge ratio – are considerably more relevant to real-life currency management. There is no consensus on how to derive ex ante returns on currencies. For most fund managers, the ex ante return on a stock consists of: o the risk-free interest rate; o the equity premium (estimates of which range from 4% to 8%), modulated by the forecast β of the stock.8 Unfortunately, for currencies these three terms are not easy to define. A discussion of this point and of currency in a CAPM framework in general is included in Chapter 5. The commonly seen approaches to deriving ex ante currency returns are: o direct forecast (rather than through a forecast of β); o assumption that currency movements are essentially random walk and use zero return for currencies; o use of the ex post returns from the last period or the historical average as the forecast (ex ante) return; 23
CO2_with_index_2nd_amends.indd 23
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
o use of interest rate differentials (forward points); and o use of the risk-free interest rate (domestic/foreign). There is an interesting issue here regarding the forward return on a currency (last two points) as opposed to the forecast return of a currency (first three points). Forward returns are derived through a no-arbitrage relationship, whereas forecast returns can be formed in a variety of ways. Forward returns are “achievable”, but they are not necessarily the best forecast of the foreign-exchange rate in the same way that forward interest rates are the best forecast of interest rates. As forward returns are achievable, we generally recommend their use as the baseline for return calculation. Other forecast-based returns can be defined relative to forward returns. We include a short discussion on the interesting and important distinction between forecast and forward in Appendix 1C. 1.3.6 Nominal return versus real return For many investment managers, the nominal – or monetary – return is sufficient. However, pension funds and some insurance funds use real returns in assessing their asset/liability position, as their liabilities are usually longer-term in nature and are linked to the rate of future inflation. Some people argue that the return on foreign currencies tends to be positively correlated with domestic inflation (when domestic inflation is high a domestic currency is more likely to weaken against foreign currencies). Therefore, the argument goes, investors should retain more foreign currency exposure than is suggested by nominal risk/return measures. As stated above, the relationship between domestic inflation and foreign-currency return says, in essence, that purchasing-power parity (PPP) holds. (Chapter 4, Section 4.2 surveys both theoretical and practical aspects of PPP.) The consensus is that PPP does hold in the long run (longer than about 8–10 years) but that the exchange rate can deviate significantly from its equilibrium value, and the process of reverting to where it “should” be can take a long time. In practice, decision makers face a trade-off between accepting the short-term fluctuations caused by currencies in order to pursue long-term benefit, and managing the fluctuations at the potential cost of long-term benefit. In our opinion the balance is tilted towards the latter because: 24
CO2_with_index_2nd_amends.indd 24
23/05/2011 14:06
Introduction to Currency Risk Management and Overlay
o the benefit of managing fluctuation is easily measurable but the benefit of long-term diversification is not; o in the current environment of stable and low inflation the benefit of diversification is perhaps not as appealing as before; and o transparency requirements and league tables accentuate, rightly or wrongly, the importance of managing short-term fluctuations. There are a handful of empirical studies on the subject,9 but their conclusions are generally dependent on the assumptions and dataset used. A theoretical debate is well beyond the scope of this book, where instead we choose to focus on issues of more practical interest: nominal return and risk measures. 1.3.7 Unhedged return, hedged return and local-currency return Unhedged and hedged returns usually refer to returns, expressed in base currency, earned on an underlying asset or a portfolio of underlying assets denominated in foreign currency or currencies. Unhedged or hedged is self-explanatory but we need to be aware of the fact that hedged return is dependent on the way hedging is done: the type and tenor of the hedging instruments and rebalancing frequency, exact timing of hedging transactions, and so on. Local-currency return usually means the return on the underlying asset expressed in local currency. In a portfolio context, or for an index, local-currency return is usually calculated simply by taking weighted average of all local-currency returns on all the constituents of the portfolio or index. In other words, local-currency return is not investable. The difference between local-currency return and hedged return on the same underlying is mainly due to the interest-rate differentials between foreign currencies and the base currency (plus some small parts attributed to factors such as rebalancing). Because interest-rate differential tends to change very slowly, the risk characteristics of local-currency return and hedged return are very similar. If you are only interested in the volatility of hedged return or its correlation with other return series and you do not have hedged return, you could use local-currency return. However, the average returns can differ substantially. One complication: some banks and data providers use “localcurrency return” to mean hedged return – always worth checking when you use external data to do analysis. 25
CO2_with_index_2nd_amends.indd 25
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
1.4 A SIMPLE DESCRIPTIVE TOUR OF CURRENCY RISKS Foreign-exchange fluctuations can have a considerable impact on the risks and returns of global investment. As shown in Figure 1.7, over any given short period (up to three years), it is not atypical to observe that 40–80% of the total risk on a foreign bond investment is attributable to movements in exchange rates. For equity investments, foreign exchange moves often account for 5–30% of the total risk. It is, however, difficult to make a definitive statement on the long-term impact of foreign exchange return on the total return, as it is heavily dependent on the sample period and the base currency of choice. Some commentators refer to periods during which the US dollar appreciated or depreciated and conclude that investors should have hedged or stayed unhedged. This kind of remark will not provide guidance for a fund manager making decisions in real life, as no one has the luxury of peering into the future with a rear-view mirror. The question of the long-term impact of currency on returns, defined in nominal terms, is the same as asking whether the impact of currency washes out in the long term. There is a considerable volume of evidence to show that it probably does, albeit over a very long horizon (the impact having a half-life of about three to five years). In this section we look at the implications of currency risk for international investment from several different angles: o short-term versus long-term contribution; o correlations between foreign exchange and the underlying; ando bonds versus equities. 1.4.1 Short-term versus long-term contribution The two panels in Figure 1.3 show the MSCI EAFE Index10 in US dollar and Japanese yen, both hedged and unhedged. We see that, over a long period, hedged and unhedged indexes follow a similar pattern, indicating that the main driver is the underlying equity returns. On the question of whether hedging is beneficial in terms of total return, we can at best answer that it depends on the period and the base currency you choose to consider. Also, it is evident from Figure 1.4 that there may be long periods of time during which the unhedged index underperforms its hedged counterpart and vice versa. 26
CO2_with_index_2nd_amends.indd 26
23/05/2011 14:06
Introduction to Currency Risk Management and Overlay
Figure 1.3 Hedged and unhedged MSCI EAFE total-return indexes in US dollar and Japanese yen MSCI EAFE (hedged into USD) 350
MSCI EAFE (USD)
MSCI EAFE (USD Hedged)
300
Price index
250
200
150
100
50
1987
1989
1991
1993
1995
1997
1999
2001
2003
2005
2007
2009
2005
2007
2009
MSCI EAFE (hedged into JPY) 350
MSCI EAFE (JPY)
MSCI EAFE (JPY Hedged)
300
Price index
250
200
150
100
50
1987
1989
1991
1993
1995
1997
1999
2001
2003
Source: MSCI, Bloomberg
27
CO2_with_index_2nd_amends.indd 27
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
In Figure 1.4 we show the two-year rolling volatility series computed from the returns series displayed in Figure 1.3.
Figure 1.4 Two-year rolling volatilities computed from the MSCI EAFE returns series from Figure 1.3
Two year rolling volatility
Annualised standard deviation
35%
MSCI EAFE (USD)
MSCI EAFE (USD Hedged)
30% 25% 20% 15% 10% 5% 0%
1989
1991
1993
1995
1997
1999
2001
2003
2005
2007
2009
2005
2007
2009
Two year rolling volatility
Annualised standard deviation
35%
MSCI EAFE (JPY)
MSCI EAFE (JPY Hedged)
30% 25% 20% 15% 10% 5% 0%
1989
1991
1993
1995
1997
1999
2001
2003
Source: MSCI, Bloomberg. Calculation by the author.
28
CO2_with_index_2nd_amends.indd 28
23/05/2011 14:06
Introduction to Currency Risk Management and Overlay
The volatility of the hedged indexes closely tracks that of the local index as currency fluctuations are removed from both series. In contrast to total returns, where we cannot make a general observation on whether hedging is beneficial, we can see clearly that currency hedging generally reduces portfolio volatility. We also notice that risk reduction is not constant over time and that there are some periods when the risk of hedged portfolios is as great as or even greater than that of unhedged portfolios. There are many academic and practical studies of currency risk as a proportion of overall portfolio risk and they typically report similar levels of risk reduction. Eun and Resnick (1994) separate the decomposed portfolio variance into four parts: 1. variance of the underlying investment in local currencies; 2. variance of the exchange-rate moves; 3. covariance of the above; and 4. cross-product terms (negligible). They looked at US-based investors investing in the French, German, Japanese, Swiss and UK equities markets and found that the averages of the four terms above, expressed as the relative contribution to total portfolio variance, are around 63%, 34%, 1% and 2%. The risk reduction achieved from hedging (removing the second and third parts) is therefore around 1/3 on average in variance terms. They calculated these figures using data from 1978 to 1994, but they are similar to later calculations by, for example, Larsen and Resnick (2000) and Clarke and Kritzman (1996). The latter presented a similar calculation for a range of base currencies. For some base currencies, fully hedging currency risk could result in higher total portfolio volatility over certain periods. The two panels in Figure 1.5 show JP Morgan global government bond indexes in both US dollar and Japanese yen, hedged versus unhedged. We will not make a comment on the total return levels between the two lines in each graph, but in general, for different base currencies and over different time periods, there is no discernible pattern for us to favour hedged or unhedged alternatives – the big return difference between hedged and unhedged indexes can be contributed to the “forward bias”, a topic we will return to in Chapter 4. 29
CO2_with_index_2nd_amends.indd 29
23/05/2011 14:06
CURRENCY OVERLAY – SECOND EDITION
Figure 1.5 Unhedged and hedged JP Morgan global government bond index in US dollar and Japanese yen Government Bond Index Global (USD Unhedged)
350
Government Bond Index Global (USD Hedged)
Price index
300
250
200
150
100
50 1993
1995
1997
1999
2001
2003
2005
2007
2009
Government Bond Index Global (JPY Unhedged)
300
Government Bond Index Global (JPY Hedged)
Price index
250
200
150
100
50 1993
1995
1997
1999
2001
2003
2005
2007
2009
Source: JP Morgan
30
CO2_with_index_2nd_amends.indd 30
23/05/2011 14:07
Introduction to Currency Risk Management and Overlay
In Figure 1.6 we show the computed two-year rolling volatility series for the returns series in Figure 1.5. Figure 1.6 Two-year rolling volatilities computed from JP Morgan government bond return series from Figure 1.5 Government Bond Index Global (USD Unhedged)
Annualised standard deviation
14%
Government Bond Index Global (USD Hedged)
12%
10%
8%
6%
4%
2%
0% 1995
1997
1999
2001
2003
2005
2007
2009
Government Bond Index Global (JPY Unhedged) Government Bond Index Global (JPY Hedged)
16%
Annualised standard deviation
14% 12% 10% 8% 6% 4% 2% 0% 1995
1997
1999
2001
2003
2005
2007
2009
Source: JP Morgan; calculation by the author
31
CO2_with_index_2nd_amends.indd 31
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
Compared with Figure 1.5, the risk reduction is much more pronounced for bond portfolios than for equity portfolios. The risk reduction is fairly consistent over the period. Again, as Eun and Resnick (1994) calculated, for a US-based investor investing into French, German, Japanese, Swiss and UK equities markets, the averages of the four decomposed variance terms listed earlier, expressed as the relative contribution to total portfolio variance, are 16%, 66%, 17% and 1%. Hedging would remove over 80% of portfolio variance on average. Similar calculations were routinely done in many studies and the results are fairly similar. More up-to-date calculations (with similar results) can be found in a recent research paper by Normand et al (2010) from JP Morgan. It calculated the risk differentials for many different base currencies. It found that the cumulative return impact from foreign-exchange moves may be modest; its volatility impact can be significant. For example, to US dollar-based investors with equity investments, risk reduction stood at 7–8% for Australian and Canadian exposure, 3% for UK exposure and 1% for Japanese exposure; it was not significant for euro exposure. To US dollar-based investors with bond investments, risk reductions were 7% for JGB, 6% for euro exposure and around 4% for UK, Canadian and Australian bond exposures. But, on a portfolio basis, US dollar-based investor would have achieved only marginal reduction in their overseas equity investment. Figure 1.7 is a summary of several similar studies. You can see that risk reduction is significant and consistent for all bond portfolios. For equity portfolios, there is usually some risk reduction but it is not consistent across base currencies or time frames. It is worthwhile to point out the limitations of these types of analysis, useful as they are. o They tend to look only at the foreign assets alone, sometime just one asset class in a single currency, and therefore they paint only a partial picture. An investor is more likely to look at both domestic assets and foreign assets, possibly across several asset classes. o They tend to show only two options in terms of benchmark hedge ratio: unhedged and fully hedged. An investor is likely to choose something between these extremes. o They tend to show volatility as the only risk measure. An investor is equally likely to be interested in some other risk measures, especially those reflecting the tail risks. 32
CO2_with_index_2nd_amends.indd 32
23/05/2011 14:07
Introduction to Currency Risk Management and Overlay
Figure 1.7(a) Risk decomposition on foreign investment portfolios: global equity portfolios
Source: JP Morgan, Goldman Sachs, Schmittman, Record Currency Management
Figure 1.7(b) Risk decomposition on foreign investment portfolios: global bond portfolios
Source: JP Morgan, Goldman Sachs, Schmittman, Record Currency Management
33
CO2_with_index_2nd_amends.indd 33
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
Note: The white bars represent the unhedged portfolio volatility and coloured bars, hedge. The difference between them can be attributed to currencies (assuming a 100% benchmark hedge ratio). If a coloured bar is taller that the white bar, it means hedging increased the portfolio risk. JP Morgan study covered 1989–2009 for equities and 1993–2009 for bonds, from Normand et al (2010); Goldman Sachs, 1990–2005 for equities and 1991–2005 for bonds, from Huttman (2005); Schmittmann (2010), covered 1975–2009; and Record (2002) covered 1980–2002 for equities and 1985–2005 for bonds.
The nominal-return-based approach, however, is not without criticism. Throughout the book, we use the nominal returns to compute the risk measures. It can be argued, however, that the investor’s risk is really to the real return rather than to the nominal return. Using this argument and a long-term data series (going back 200 years!), Froot (1993) looked at the impact of hedging in the long term. He used the real returns on US-dollar-denominated assets in the domestic currency (in this case, sterling), the real return being the US dollar return adjusted for spot foreign-exchange movements and inflation in the UK. He argued that the real exchange rate is mean-reverting over the long term, and this observation is borne out by his long data series, albeit for only one currency pair. When the real exchange rate is mean-reverting, currency hedging is not effective in the long term. Hedging introduces interest-rate differentials to the long-term performance, which may even increase the portfolio variance. Froot concluded that, although hedging is effective in the short term, its effectiveness ceases beyond five years for equities and eight years for bonds. He also noted that the minimum-variance hedge ratio declines from close to 100%, as argued by Perold and Shulman (1988), to an average of about 35% over a horizon of five to ten years. He argued that, if short-term currency fluctuation is not a major concern for investors, they should not hedge currency simply to avoid transaction costs. Froot’s (1993) data series covers several currency regimes, including the old gold and silver (bimetallic) standard, the gold standard and the current floating-rate regime, so we may question the particular relevance of his results for the current foreign-exchange environment. Also, the method overlooks the risk-reducing11 and return-enhancing benefits that a sound currency-hedging programme can bring. In a more recent update incorporating more base currencies and exposure currencies, but not quite as far back in history, Schmittmann (2010) found hedging remains effective for horizon as long as five years. 34
CO2_with_index_2nd_amends.indd 34
23/05/2011 14:07
Introduction to Currency Risk Management and Overlay
1.4.2 Correlation The correlations between currencies and the underlying asset returns have been investigated both theoretically and empirically. As far as we know, there is no consensus on how such correlation should be properly defined. Quite a few studies we have seen calculate the correlation between currency returns and foreign asset returns denominated in the domestic currency. It is difficult to interpret the correlation coefficients obtained in this manner, as the latter already contains the exchange-rate element. Our own recommendation is to calculate correlation using exchange-rate returns, spot-on-forward and hedged underlying returns. This may be a more cumbersome calculation, but it is more suited to the task of currency risk management. The calculation can be simplified by correlating exchange-rate returns, spot-on-spot and underlying returns denominated in local currencies. As long as interest-rate differentials are not large and/or are reasonably constant, the calculated results will be similar. Many people comment that the correlation between bonds and currencies is high, whereas that between equities and currencies is low. If we calculate correlation coefficients using domestic returns and currency returns, the above conclusion is the natural result. This is because currency volatility is usually higher than bond volatility, so returns on foreign bonds denominated in the domestic currency will be predominantly currency fluctuations. For equities, the effect is much less dramatic simply because equity volatility is typically higher than currency volatility. So correlation calculated in this way does not offer any insight into whether there are intrinsic links between currencies and the underlying, or on whether currency fluctuations should be removed. We will comment further on the correlation between currencies and bonds or equities in the next two sections. It is a well-recognised feature in currency risk management that correlations between currencies and underlying assets are not constant. In a paper, Goetzmann, Li and Rouwenhorst (2001) show that the diversification benefits of global investing are not constant, and that these benefits are currently low compared with other periods in the history of capital markets. The authors decompose diversification benefits into two parts: a component that arises from variation in the average correlation across different national markets; 35
CO2_with_index_2nd_amends.indd 35
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
and a component due to variation in the investor opportunity set. There are periods, such as the last two decades, in which the opportunity set expands dramatically and the benefits of diversification are driven primarily by the existence of marginal markets. In other periods, such as World War One and the two following decades, risk reduction is due to low correlations among the major national markets. They infer that periods of globalisation have both benefits and drawbacks for the international investor. Globalisation expands the opportunity set, but diversification relies increasingly on investment in emerging markets. They conclude that the global correlation structure is not constant between various periods. Normand et al (2010) observed that, in general, the covariance of currency returns with asset returns must be sufficiently negative to overcome the volatility of the currency itself. The instability in the correlation between currencies and the underlying assets undermines the validity of many analyses in currency management. For example, when we use CAPM optimisation to determine optimal currency weights or to find the optimal hedge ratio, we have to be extra careful, as the results may vary significantly when different sets of correlations between currencies and the underlying assets are used. This is a topic we will return to in more detail in Chapter 5. 1.4.3 Stock prices and currency fluctuations It has generally been assumed that the correlation between stock or equity prices and currency moves is zero or close to zero. The assumption enabled us to equate country exposure to currency exposure and simplified the task of quantifying currency risk because, historically, global equity management has been structured around country allocation. Country allocations were first chosen and the stock picking was then conducted within countries. The currency risk associated with a particular shareholding was fully allocated to the corporation’s headquarters or trade listing location. However, over the last 10 to 20 years, the environment in which corporations operate has undergone considerable change and the corporations have reacted accordingly, primarily by going global themselves. These days, many corporations have to be viewed and valued as a portfolio of activities across different currencies. How36
CO2_with_index_2nd_amends.indd 36
23/05/2011 14:07
Introduction to Currency Risk Management and Overlay
ever, matters are not as straightforward here as they might appear, and a simple breakdown of a corporation’s revenue into different currencies will not necessarily reveal the true extent of its currency exposure. A range of factors come into play: o many corporations actively manage their exposure to currencies through various financial instruments and through the currency mix of their debt; o many corporations maintain excess capacity in different countries as an operational hedge – they can shift production from one location to another in response to persistent currency moves; o some corporations have pricing power that enables them to pass on exchange fluctuation to consumers; o in some situations corporations can be indirectly exposed to exchange movement if their competitors are exposed; o some corporations have changed their functional currencies.12 It could be argued that every corporation merits a separate model that delves into these factors, but in a portfolio-management context this is clearly undesirable. We think that the market-based approach proposed by Solnik (1993) offers a valid and practical alternative to investigate the issue. He assumed that equities are priced globally with currency risk taken into account. Therefore, if we regress the historical returns of a particular stock on the domestic index return, foreign index returns and the movement of foreign currencies against the domestic currency, we can obtain the sensitivity of the stock’s returns to all the factors – especially to currency moves. These sensitivity estimates are called “loadings”, which in optionpricing terminology are equivalent to deltas.13 A holding in a particular stock can then be decomposed into different currency-sensitive parts, which can be aggregated across a portfolio. The problem with this approach is that a corporation’s exposure to currencies can change dramatically following a particular corporation action. Jorion (1990) suggested that an alternative measure of foreign currency exposure can be obtained by regressing individual firm stock returns against the market-return and exchange-rate movements. This measure of the sensitivity of stock returns to exchange rates provides a net-of-hedging measure of foreign currency exposure, since market returns should already take into account any 37
CO2_with_index_2nd_amends.indd 37
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
hedging activities undertaken by the firm. This approach has been applied to various datasets by Amihud (1994), Bartov and Bodnar (1994), Donnelly and Sheehy (1996) and Kedia and Mozumdar (1999), extended to multiple exchange rates by Miller and Reuer (1998), and applied to different return horizons by Chow, Lee and Solt (1997). Of these studies, only the UK study by Donnelly and Sheehy found statistically significant results for the sensitivity of stock returns to exchange rate changes, which they attributed to the relative openness of the UK economy. Although the Jorion measure is a theoretically appealing measure of foreign currency exposure, the existing empirical evidence indicates that the exposure coefficients are highly volatile and not generally related to foreign operations (Kedia and Mozumdar 1999). A practical assumption on the correlation between hedged equities and currency returns is that it is zero. If we calculate the correlation coefficients over a reasonably long period, they are usually small and not significantly different from zero. 1.4.4 Bond prices and currency fluctuations Several academic studies look into the risk factors that determine bond returns,14 but very few have examined the relationship between currencies and international bond returns. Driessen, Melenberg and Nijman (2000) used weekly bond index data from 1990 to 1999 for Germany, the US and Japan and principal component analysis to detect the factors that impact on bond returns. They used both currency-hedged and unhedged indexes, and their results shed light on the question of bond prices and currency fluctuation. They found that currency fluctuations do seem to explain a portion of the variability of unhedged bond returns (about 25%); however, they concluded that there is no reward in bearing currency risk. Annaert (1995) found that exchange-rate fluctuations imply higher correlations between returns, and thus higher estimation risk. He also computed the correlations between local bond returns and exchange-rate returns for investors based in the US, Germany and Japan investing in a range of bond portfolios in 15 countries. They are usually very low and not significant. Of 42 such correlation coefficients he reported, only seven were significant at the 1% level, most of them falling between −10% and +10%. 38
CO2_with_index_2nd_amends.indd 38
23/05/2011 14:07
Introduction to Currency Risk Management and Overlay
Again, a practical assumption on the correlation between hedged bonds and currency returns is that it is zero, but this may be more controversial. If we calculate the correlation coefficients over a reasonably long period, as for hedged equities, they are small and not significantly different from zero. However, hedging decisions on bond portfolios are usually made on strategic grounds simply because of the relative magnitude of currency volatilities, which renders this sort of agonising assumption unnecessary. 1.4.5 A word of caution for unsuspecting readers of statistics We have avoided being overzealous in the above sections in terms of presenting table after table of returns, volatilities, correlations, optimised hedge ratios, efficient frontiers and so on. Plenty of these can be found in the references cited in this section, and in the marketing materials provided by overlay managers and other involved parties. One reason for their frequent use is that this type of analysis is increasingly easy to conduct with the availability of data and computing capabilities. An important reason for our caution here is that summary statistics have some serious shortcomings. The historical average returns of hedged versus unhedged study tend to be very noisy: choosing a different base currency or a slightly different time window, the average returns could change dramatically. Risk measures such as volatilities and correlation coefficients may gloss over vital information and lead to erroneous conclusions. For example, both are informative measures for normal or near-normal distributions. However, the usefulness can break down when we move away from normal distributions – especially that of correlation coefficients, which can be heavily influenced by a handful of outliers. Statistical remedies abound, but they can quickly move into realms beyond a busy practitioner’s will or ability to reach. One remedy we invariably find useful and reliable is a chart. Before making a decision based on computed correlation coefficients, it is always helpful to take a quick look at a scatter plot of the correlated variables to see if there are data points that are way out of line with the rest. This is only common sense, but more often than not we find it in preciously short supply. Here are a few practical tips on interpreting summary statistics usually included in hedging analysis. 39
CO2_with_index_2nd_amends.indd 39
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
o The average returns are the least reliable in general, with perhaps only two exceptions. They are unreliable not because they are calculated wrongly, but they offer very little information on what kind of average returns you might get in the future. The two exceptions are the long-term returns that can be attributed to carry trades and returns from emerging-market currencies. They are somewhat more reliable. o Average volatilities over a medium to long period are reasonably reliable in that the future volatilities are likely to revolve around those averages. However, average volatility tends to underrepresent the large, damaging moves that are probably more important for risk-management purpose. o Average correlations between currency and underlying asset returns will only be somewhat reliable if it is over a long period. But, over a long period, they tend to be indistinguishable from zero. o Optimised hedge ratios are unreliable if historical returns are used in the optimisation; minimum-variance optimisation is slightly better but it suffers from unstable correlations. o Efficient frontiers are useful only in terms of graphically summarising information. 1.5 WHAT IS SPECIAL ABOUT CURRENCIES? There are various reasons why currencies may be considered special in the context of international investment. These reasons are often linked to something known as the “currency hedging debate”, which revolves around the question whether an investor who invests in foreign-currency-denominated assets should hedge the embedded currency risk.15 The question is as old as global investment itself. While the idea of spreading your investment globally has gained general acceptance – if with the occasional dissenting voice – 30 years of debate on the question of hedging the associated currency risk has not produced a consensus. Far from it. Sometimes the currency-hedging debate is formulated in a rather more specific way: “Should an investor (passively) hedge 0%, 100%, 50% or 77% of the currency exposure associated with international investment?”16 Another related, but slightly different, question is whether active currency management can add value. Investigation in this field has followed two routes, one looking at the predictability of 40
CO2_with_index_2nd_amends.indd 40
23/05/2011 14:07
Introduction to Currency Risk Management and Overlay
currency movements (a topic we review in Chapter 4), and the other examining the historical performance of active currency managers (the evidence on this will be considered in Chapter 2). In the following sections we take a brief look at some of the statements that usually feature in (and, more often than not, spoil) the currency hedging debate. Our emphasis here is not on demonstrating that a particular statement is right or wrong: many of the questions are, in our view, open-ended and are likely to remain so. A definitive answer to any of these questions may require a number of assumptions and restrictions, resulting at times in a quagmire of arguments and counterarguments. We instead emphasise the point that, irrespective of your answer, currency risk in a portfolio is real and needs to be managed. The currency-hedging debate itself will be discussed in Section 1.6.3 and empirical evidence relating to the usefulness of currency hedging is reviewed in Section 1.6.4. 1.5.1 “Currency trading is a zero-sum game” The argument that currency trading is a zero-sum game comes in two related forms. 1. One market participant’s currency gain is another one’s loss and therefore currency trading is a zero-sum game. Taking transaction costs into consideration, it is probably a negative-sum game. 2. In the long run, currency impact washes out: you gain one day (or year) and you lose another. The first point is probably correct if we define gain and loss in pure cash terms over a very short time horizon. For a given buyer X, there will be a seller Y on the other side, and one’s gain is clearly the other’s loss. In practice, many factors come into play when gains and losses are defined. This can be done in a variety of ways – for example, relative to certain benchmarks, short-term or long-term, in percentage terms rather than cash terms, or with or without adjustment for risk, including or excluding the returns that could be earned on the back of currency transactions, etc. The sum of the gains and losses defined in different ways may not be zero.17 The second point is probably correct too if we are looking at a horizon as long as, say, 8–10 years, or longer, and if we include cases where investors’ base currencies 41
CO2_with_index_2nd_amends.indd 41
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
are different. For shorter horizons, however, currency fluctuation is significant, and for certain base currencies the historical returns may have been consistently positive or negative. If we assume that both assertions are correct, can we conclude that currency risk should not be managed? We do not think so. Not everyone in the currency market is profit-driven. For example, central banks and corporations are active in the market to dampen volatility and to hedge their exposures, respectively. Hence, they seek non-monetary “gains” from trading in the currency market. In fact, only a very small proportion of the gigantic currency transaction volume is there purely for profit motives. It is therefore possible to trade currencies for excess returns, ie, the returns in excess of risk-free interest rates. One major motivation for managing currency risk is to make people accountable for the risks they take. Even if currency movements average out in the long-term, there are still plenty of short-term fluctuations. Although some still maintain that these can be disregarded if we accept the long-term zero-return argument, they increasingly impact on real performances through peer-group ranking and other performance-measurement practices. 1.5.2 “Currency movements represent random walks and are unpredictable” A wealth of literature documents the difficulties of predicting exchange rates using structural or time-series models. Frankel and Rose in their well-regarded 1995 survey on the subject concluded that traditional macroeconomic models of foreign-exchange determination have relatively little explanatory power. They also expressed doubt about the value of further time-series modelling of exchange rates at high or medium frequencies using fundamental models. However, the research effort in the area has shown no sign of slowing down since their study. On balance, the body of evidence seems to support the claim that currency moves can be predicted, if only partially. Many of the later models focus on quantitative methods, especially those taking nonlinear features of exchange-rate movements into account. The line between technical analysis and quantitative analysis has become much finer thanks to the availability of data and computing power. The quest for the 42
CO2_with_index_2nd_amends.indd 42
23/05/2011 14:07
Introduction to Currency Risk Management and Overlay
“Holy Grail” continues, and we discuss foreign-exchange prediction models in Chapter 4. Again, the predictability question is not the same as the currency risk management question: an investment manager who believes that currency moves are intrinsically unpredictable – and there is plenty of evidence to back up this view too – still faces currency risks. The latter are quantifiable (and possibly predictable too – see Chapter 8), and we firmly believe they should be measured and taken into account. 1.5.3 “Foreign exchange is not a separate asset class” There is no consensus on whether foreign exchange is a separate asset class. We think that it is probably not in the traditional sense. Without being too technical, a type of asset can be considered a separate asset class if: o the returns on the asset are not highly correlated to returns on existing asset classes or their combinations; o the returns on the asset are correlated to known systematic risk factors, for example, market access return, inflation, changes in industrial productions, etc.;18 o there is a risk premium for assuming these systematic risks; and o capital is needed to invest in such asset. Here we need to distinguish passive currency returns from active currency returns. The former refers to the returns from a single or a portfolio currency position, where the positions are held passively; whereas the latter refers to the returns from an actively managed currency position or a portfolio of positions. Passive currency returns are usually not highly correlated to the returns of other asset classes. However, the explanatory power of commonly used risk factors is limited when applied to currency movements; there is no evidence that a stable currency risk premium exists; and capital is not necessary to establish currency positions. Hence our view that currency is not a separate asset class in the traditional sense. Currency is, however, viewed by many investment managers as a “tactical asset class” (equities and bonds and real estates and so on are “strategic asset classes”) for the following reasons:
43
CO2_with_index_2nd_amends.indd 43
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
o the correlation between active currency returns and those from strategic assets is low (several studies over recent years seem to support this point – we will revisit this point in Chapter 2, Section 2.8.3); o projected active returns are positive (some supporting evidence has been presented in recent years, though not without qualification; this is also reviewed in Chapter 2, Section 2.8); and o capital requirements are low (generally true). If all these are true, a mean–variance portfolio optimisation model will allocate capital to such tactical assets. The distinction between “tactical asset class” and “strategic asset class” is no longer a debating point these days, as most investors accept the hedge fund as a distinctive asset class in that they have specific allocations to the hedge fund strategies. A hedge fund is not really a “strategic asset class”: rather, it is a “managed alpha asset”, or just “alpha asset” – the old “strategic asset” becomes “beta asset” due to CAPM pricing relationships. If an investor treats a hedge fund as a separate asset class, then active currency management should be a separate asset class too – the returns from hedge funds as well as from active currency management are dependent on managers’ ability to trade. In fact, the distinction between one of the established hedge fund strategies, “global macro”, and some of the active currencymanagement strategies, is quite blurred. Also, the answer to the question whether currency should be treated as a separate asset class does not need to be the same as a view on whether currency risk should be managed. Currency risk is an avoidable consequence of global investment, irrespective of whether or not currency is viewed as a separate asset class or whether excess return can be generated from it.19 Since it is a risk factor, it should be managed under the proposed risk-measurement and risk-accountability framework. Many asset managers adopt a pragmatic approach to the question of whether currency constitutes a separate risk class. For example, UBS Global Asset Management has used the classification of its investment categories given in Table 1.3. Currencies are listed in the opportunistic investment category. Some asset managers may consider currencies as alpha-intensive (ie, not quite an asset class in the 44
CO2_with_index_2nd_amends.indd 44
23/05/2011 14:07
Introduction to Currency Risk Management and Overlay
traditional sense but with a high likelihood of consistently yielding excess returns) or even include currencies in their core portfolios. Table 1.3 UBS Global Asset Management’s investment categories Core portfolio
Alpha-intensive
Opportunistic
Global equities
Hedge funds
TIPS
Developed & emerging markets
Global natural resources
Sector funds
Global fixed income
Mortgages
Developed & emerging markets
Corporate debt REITs
US high-yield bonds Foreign real estate Private equities Currencies Real estate equity Dynamic trading strategies Source: UBS Global Asset Management, 2001
1.5.4 “Currency is part of global diversification, but hedging reduces diversification benefit” It is typically claimed that unhedged international equities give better diversification than hedged equities because the correlation between domestic equities and hedged foreign equities is generally higher than if the foreign equities are unhedged. However, this argument is flawed as the reduction in correlation is expected if currency returns are not correlated with the equity returns. Generally speaking, if we add a random variable (in this case, currency returns) to one of the correlated variables (domestic and foreign equity returns), the correlation among the latter will be reduced. The 45
CO2_with_index_2nd_amends.indd 45
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
higher the currency volatility, the more the correlation is reduced. If currency volatility is very large relative to equity volatility, the correlation between unhedged foreign equity returns and domestic equity returns can approach zero as the “true correlation” becomes drowned by the noise produced by currency movements. And this is more or less the case for international bond portfolios. Currency volatility generally dwarfs bond volatility, so returns from unhedged foreign bonds typically exhibit low correlation with those from domestic bonds. Therefore, from the fact that hedged foreign asset returns tend to show higher correlation to domestic asset returns compared with unhedged foreign asset returns, we cannot conclude that hedging is detrimental or beneficial – if a reduction of correlation is good on its own, then adding horse betting would have been a valid investment strategy. The impact of currency hedging should be measured by comparing the hedged portfolio risk with the unhedged portfolio risk. On this measure, the empirical evidence, including that surveyed in Section 1.6.4, suggests that hedging can reduce portfolio risk. But, again, this is not without controversy, as the conclusion depends on how we measure portfolio risk and what assumptions are used for correlations, etc. There is, however, a deeper and more important angle to the issue of diversification: a certain amount of currency exposures may be good as a hedge to domestic inflation. When domestic inflation goes up – bad news for domestic pension funds and insurance companies, since their liabilities are likely to rise – the domestic currency tends to go down. Having some unhedged foreign currency exposures therefore makes sense. Some market commentators suggested a formula for the optimal percentage of foreign-currency exposures based on the idea of inflation hedge, which we will review in Chapter 5, Section 5.6. The empirical evidence on this claim is inconclusive. Some commentators argue that currency hedging can reduce the portfolio volatility without reducing the portfolio returns, and therefore represents a “free-lunch”. 1.5.5 “Unlike equity or bonds, currency does not attract a risk premium” Both equities and bonds attract a risk premium, which is the return in excess of risk-free interest rates. For currencies, the risk pre46
CO2_with_index_2nd_amends.indd 46
23/05/2011 14:07
Introduction to Currency Risk Management and Overlay
mium is not clearly defined. Historical analysis conducted along these lines has produced inconclusive results – after all, we have meaningful foreign-exchange data for less than three decades, and this is deemed insufficient to conduct a useful significance test.20 Using what is available and some fancy statistical tools, some conclude that emerging-market currencies attract a risk premium (before they experience a landslide devaluation, that is!21). As for the currencies of developed economies, there seems to be no evidence to suggest that taking systematically long or short positions will be rewarded by a risk premium. But then there are a host of phenomena (termed “puzzles” in the financial literature) that seem to contradict the zero-premium assumption, and these phenomena are usually linked to the unexplained profits/losses from holding currency positions in a particular manner – one of them being forward bias (discussed in the next subsection). The current consensus seems to be that there probably is a currency risk premium but that it is highly unstable and can be either positive or negative. Capital-asset-pricing-model (CAPM) theory incorporating currencies links currency-risk premium to factors such as the aggregate risk-aversion of investors based in different countries and the net asset positions of countries, among others. This point will be discussed again in Chapter 5, Section 5.2, where the topic of currency and CAPM is reviewed. However, the practical relevance of whether there is a risk premium for currencies is probably limited for a currency manager. Some argue that, since forward contracts are in zero net supply and gains by some counterparties are offset by losses for the others, currency offers no risk premium. But this argument is flawed. We can think of stock index futures that are also in zero net supply where, nevertheless, it is generally accepted that equity investment attracts a risk premium. 1.5.6 “Currency market is efficient” Here the argument against currency hedging says that, because the currency market is very efficient, there is no point in trying to beat the market price, which summarises all available information. Such attempts are deemed fruitless. The foreign-exchange market is always commented on for its 47
CO2_with_index_2nd_amends.indd 47
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
huge volume and liquidity. However, there is no consensus on the efficiency of this market. Chapter 4 reviews several models used by participants that are said to exploit inefficiencies in the foreignexchange market. The debate on whether the efficient-market hypothesis is applicable to the foreign-exchange market is not confined to empirical matters. Many commentators question the assumptions that have to be made to ensure the validity of the efficient-market hypothesis. It assumes that investors (a) are rational and (b) that they aim to maximise profit. However, numerous studies, both theoretical and empirical, demonstrate that investors may not be rational. In fact, it is common to observe herding behaviour among investors (when they are often known as “noise traders”). Many researchers attribute excessive short-term fluctuations in exchange rates to such herding behaviour.22 The noise traders may overreact to news and generate exchange-rate movements that have little relation to fundamentals. In this case exchange rates may be reacting to noise rather than news, and movements could simply induce further movements through a feedback mechanism. Also, the assumption about the profit-maximising objective of market participants may not apply to all – for example, to central banks, whose mandates generally require them to dampen fluctuation rather than seek profit. It is possible that this type of trading activity can generate patterns in exchange rates and present arbitrage opportunities. The efficiency of the foreign-exchange market is a wide-ranging question that relates to many of the points discussed in this section. Any answer to the question will probably have to be balanced with a number of qualifications and, in our opinion, may not offer much guidance on what to do about currency risk. We shall therefore maintain a pragmatic approach and focus on risk measurement and accountability. 1.5.7 Forward bias (uncovered interest-rate parity, or carry trade) This and the next two subsections are more about “what’s special about currency” rather than the currency-hedging debate itself – we group these topics together here because they are usually discussed together in a currency-hedging context. Currencies trade in both spot and forward markets. The dif48
CO2_with_index_2nd_amends.indd 48
23/05/2011 14:07
Introduction to Currency Risk Management and Overlay
ference between spot and forward prices, the “forward point”, is determined by the interest-rate differential between the currencies. The forward price is derived from an arbitrage relationship rather than from some consensus expectation in the market (as an example of the latter, forward interest rates are based on a market consensus expectation of where interest rates will be in the future). This is an interesting and very important distinction, and we will expand on it in Appendix 1C. Nevertheless, to avoid arbitrage, currency spot prices should converge to the corresponding forward prices. But in general they do not. This is so-called “forward bias” and is frequently quoted as evidence of inefficiency in the currency market and as one of the major puzzles of international finance. Forward bias is widely used as the basis for currency-trading strategies, which is typically referred to as “carry trade”. Various explanations for the effectiveness of such strategies have been offered, for example: o Time-varying risk premiums. As the forward price rate may incorporate a risk premium, the spot price will not converge to the forward price even if the market is efficient. o Misspecification of the expectation. In this view, the spot is still expected to converge to the forward price, but the way we define the expectation is wrong. For example, there might be a potentially disastrous scenario that is priced in the forward but has not yet happened (the “peso problem” again). Therefore a test result relying on the historical data cannot capture this latent move – although, if we had a sufficiently large dataset, the problem would disappear. Forward bias is a very important topic and has been thoroughly investigated in both the theoretical and practical literature – it has received wide publicity over recent years due to the supposedly large speculative positions following this strategy. The collapse of the Icelandic krona – along with, but to a lesser extent than, some other high-yielding currencies such as sterling and the Australian dollar – during the global financial crisis was generally blamed on the unwinding of carry trade positions. In our view, forward bias still represents a valid model for active trading, probably not in its naïve form. However, this is not a reason to ignore currency risk. 49
CO2_with_index_2nd_amends.indd 49
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
If anything, profit made through forward bias-based strategies should also be evaluated against the risk involved. The topic of forward bias will be revisited in more detail in Chapter 4, Section 4.5.1. 1.5.8 Autocorrelation (momentum, trends and countertrends) Empirical evidence suggests, though not conclusively, that currency returns show positive autocorrelation over longer horizons (namely, an up move is more likely to be followed by another up move) and negative autocorrelation at a tick-by-tick level (an up tick is more likely to be followed by a down tick). The former is widely exploited by market participants, and there are countless articles and books on the subject. A short overview is provided in Chapter 4, Section 4.5.2. The latter has been investigated empirically only in recent years with the availability of data and the computer power to handle large databases. Is the existing evidence enough to provide a basis for taking trading decisions? Probably. One indirect proof is simply that the majority of active players in the currency-trading arena can be classified as trend players of sorts. While the trend-following models are growing in sophistication, there is, however, growing evidence to indicate that the returns they give have been dwindling over recent years. The positive autocorrelation over longer horizons has been attributed to the herding behaviour of market participants and the self-fulfilling prophesies of trend following, among other things. The tick-level negative autocorrelation is said to be linked to microstructural features. o Market-makers’ order books are skewed in different directions and they will skew their prices accordingly. Actual prices may touch different sides of the bid–offer spread from different banks in succession (ie, a buying transaction from one bank is more likely to be followed by a selling transaction from another bank). o Heterogeneous traders react to a piece of news differently. 1.5.9 Purchasing-power parity (PPP) prevails in the long run The concept of PPP goes back to an early work of Cassel (1916) and simply measures the similarity of consumption opportunities 50
CO2_with_index_2nd_amends.indd 50
23/05/2011 14:07
Introduction to Currency Risk Management and Overlay
in different countries. Following the law of one price, similar consumption baskets in different countries should have the same price tag. If not, exchange rates should adjust until they do. In international finance there are two versions of PPP: o absolute PPP: the exchange rate between the currencies of two countries should be equal to the ratio of the average price levels in the two countries; and o relative PPP: this focuses on the relationship between the inflation rates in two countries and the change of the exchange rate between the countries’ currencies over a certain period; if relative PPP holds, exchange rate shifts perfectly mirror inflation differentials and have no influence on the valuation of asset returns in real terms. Consumption baskets and prices are reasonably easy to measure (though not so easy with all the twists and turns dreamt up by economists and statisticians), but observed currency prices have been shown to deviate consistently from their “parity values” suggested by PPP. Important and intuitive as it is, in practice PPP has only limited value as a foreign-exchange determination model. It can be damaging to use it as a reason to reject the practice of measuring and managing currency risks. We will review PPP models in detail in Chapter 4, Section 4.2. 1.5.10 Summary Through this brief exposition of what is special about currencies, we hope to have established that a clear-cut answer (or lack of one) to any one of the questions we considered may not provide a direct answer to the currency-hedging debate. Many of the questions are linked to, if not derived from, foreign-exchange research and they will be the main subject of Chapter 4. There are two more topics to cover in considering what is special about currencies, and these are: 1. Excess volatility: macroeconomic models consistently fail to account for the level of exchange rate volatility. 2. Volatility memory (clustering): this refers to the high autocorrelation of short- and long-term volatility.23 51
CO2_with_index_2nd_amends.indd 51
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
The first will be discussed in Chapter 4, Section 4.3. The second, volatility clustering, offers opportunities for excess returns. In fact, the whole subject of currency volatility is so important that it will be reviewed in Chapter 8. 1.6 FX RISK MANAGEMENT AND CURRENCY OVERLAY 1.6.1 Continued interest in foreign-exchange risk management The question of foreign-exchange risk management in the context of global portfolio management has been around for as long as global portfolio management itself – though it really gained prominence in the new era of floating exchange rate since 1973. Academics did not ignore this question, as extensions have been made to the CAPM to incorporate foreign exchange (see Chapter 5, Section 5.2). There are also numerous empirical studies on the different aspects of foreign-exchange risk management. However, there is no agreement on how foreign-exchange risk should be managed in an international portfolio. In fact, as we have seen, there is no consensus on whether there is foreign-exchange risk in such a portfolio. In the years before the publication of the first edition of this book in 2003, there had been a renewed interest in foreign-exchange risk management from the fund-management industry. At the time, information about methods to manage currency risk was patchy and the idea of currency overlay was not as widely understood as it is now. That was partly the motivation behind the first edition. The reasons behind the increasing level of interest in currency management then can be traced to the following. o Demographic changes have resulted in an explosive growth of the global asset-management industry. o Increasingly larger proportions of assets under management (AUM) are invested globally in pursuit of higher returns and/ or diversification. o International capital flows have a substantial, and sometimes sustained, impact on exchange-rate movements as compared with traditional factors such as current-account balances. Increased capital flows are one of the factors contributing to the heightened volatility of currencies.
52
CO2_with_index_2nd_amends.indd 52
23/05/2011 14:07
Introduction to Currency Risk Management and Overlay
o Investors are paying closer attention to returns and risks on their investments by looking at league tables and other performance indicators. Currency movements can produce highly undesirable swings in performance. o A highly liquid and efficient foreign-exchange spot and options market has emerged to price, transfer and manage risk. The depth of this market and its capacity for innovation continue to break new ground. o Less expensive computing power and availability of data make risk separation, optimisation and performance attribution much easier. o New risk-measurement methodologies, such as VaR, have been developed and are gaining wide acceptance. The regulatory environment in most parts of the world is moving towards a tighter, fairer, more transparent approach. Requirements such as mark-to-market accounting and minimum funding levels compel fund managers to be more alert to currency risks. The poor performance of stock markets around the world after the burst of the dotcom bubble has also propelled investment managers into focusing on investment policy, including currency policies. In its quarterly survey, the Defined Contribution Universe Summary, published in 2003, Mercer Investment Consulting, a pension fund consultancy, stated that nearly 70% of defined-contribution plans in the US now put out an investment policy statement, compared with 57% two years ago. Many funds, in the process of clarifying the role of asset allocation and risk responsibility, have started to reconsider the question of currency in international investment and the management of currency risks. In a 2002 survey by JP Morgan Fleming Asset Management (2003) on alternative investment strategies, 9% of correspondents said they used currency overlay, 25% were considering whether to do so, and the remaining twothirds did not use it. Australia’s National Australian Bank (2005) found, in their second Australian Superannuation FX Survey, that four-fifths of superannuation funds surveyed hedged their international equity portfolios, one-fifth of respondents with benchmark hedge ratio 53
CO2_with_index_2nd_amends.indd 53
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
of 50% or more. The average benchmark hedge ratio was 35%. In addition, 56% of the superannuation funds surveyed had changed their international equity hedge benchmarks in the previous two years, while 70% had changed them in the previous five years. Most superannuation funds took into account changes in the level of the Australian dollar when setting their benchmark but may take the opposite decision: some increased hedge ratio to avoid the effect of further rises in the currency, whereas others saw the currency as being set for a fall and reduced hedge ratios. Four years and a financial crisis later, the latest survey (National Australian Bank, 2009) found that average benchmark hedge ratio had been stable, at around 45%. Some 22% of funds had a “portfolio approach” (targeting a fixed currency exposure across asset classes; see Section 3.3.2 for further discussion on this point) compared with only 5% in 2007. Eighty-five per cent of super-funds believed currency issues were important or very important, compare with 74% in 2007. Interestingly, the use of specialist currency-overlay managers for either passive or active currency management had dropped from 2005 highs of 84% to 76% – some super-funds had moved currency management in-house and smaller funds increasingly used their custodian for this service. Callan Associates, a US investment consultant, surveyed 59 mainly US-based pension funds in 2005, covering close to US$1 trillion of AUM. The survey found that most pension funds allocated more than 10% to international equity, with large public pension funds allocating 19%. Eighteen per cent of surveyed pension funds had explicit currency policies and they generally favour external overlay managers. The reason for using currency-overlay programmes is mostly risk reduction, or risk reduction plus return enhancements. Also worth nothing is that the majority of pension funds allow international equity managers to hedge opportunistically. Brandes Institute (2007) noted that, in September 2006, the top 200 US defined-benefit plans had committed US$74 billion to currency-overlay programmes. That’s around 18% of the total assets these plans have in international active equity portfolios, and a larger percentage than the amount committed by these institutions to emerging-markets equity, hedge funds or international bonds. Johnson (2009) re54
CO2_with_index_2nd_amends.indd 54
23/05/2011 14:07
Introduction to Currency Risk Management and Overlay
ported that some 6% of US institutions had adopted an active currency programme, according to research by Greenwich Associates; and in the UK the National Association of Pension Funds found 12% of its members used active currency programmes. However, Mercer estimated that some 33% of UK investors are now using some sort of currency hedging within their portfolios in 2009, up from 25% a year before. Pichardo-Allison (2009) reported that State Street’s passive overlay business was three times larger than in 2007. Meanwhile, interest in active currency management has been in decline since the global financial crisis. According to data compiled by Mercer, global searches for active currency managers has decreased since 2006 with 16 searches recorded that year, 11 in 2007 and a mere 5 in 2008. Most recent client interests had been on passive overlay business. 1.6.2 The growing global market and its consequences for investors There are plenty of books and databases on the scope and depth of the global capital market and investment, the rapid expansion of which provides the backdrop for the topic of currency management. In the first edition, we included a few charts to show that pension funds in developed countries had increased their allocations to global equities; that the global investment universe had expanded considerably; that both US assets held by foreign investors and foreign assets held by US investors had been on a steady climb. These are all still ongoing, larger, faster, and more. Compared with the situation then, it now seems superfluous to include a similar section in the new edition on the increasing trend in global diversification and its consequences for investors. We make a few observations here. o Home bias has come down over the years in nearly all developed countries.24 o Assets held by institutional investors and invested cross-border have increased dramatically.25 o Investors based in emerging-market economies have increased their foreign-currency-denominated holdings significantly, including the massive increase in reserve holdings. And more 55
CO2_with_index_2nd_amends.indd 55
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
and more are run according to commercial principles of seeking return and reducing risk rather than as government bureaus. o Corporations are more global, which may have implications on how underlying securities react to foreign exchange rate moves. The conclusion is the same: the investment universe is increasingly global and investors’ exposure to foreign assets continues to grow. Against this background, considerations of currency risk become increasingly relevant. 1.6.3 Should currency risk be hedged? Some questions and answers Whether currency risk should be hedged is a question that lies at the centre of the “currency-hedging debate”, and the answer is clearly worth considerably more than sixty-four thousand dollars! In many ways the whole book is about this question. However, we do not think we should seek simple yes-or-no answers that are applicable under different sets of objectives and constraints. In our opinion, currency risk is real and can be substantial, especially given the expansion of the global investment universe. Currency risk should be measured separately from the underlying risks. Moreover, currency risk should be entrusted to managers with comparative advantage. Nor do we necessarily propose that a currency be hedged, either passively (using a predefined hedge ratio) or actively; a fund can choose to leave currency risk unhedged as a strategic and informed decision. However, if such a decision is made, it is our belief that a fund should centralise the currency risk at the fund level and measure the performance of the underlying managers as if they were investing on a fully hedged basis. Later in the book we will provide a balanced overview on the whys and some practical techniques for the hows. Below we divide the question whether currency should be managed into several sub-questions and provide brief answers. Detailed answers are given in the relevant chapters. Q1: Does passive hedging add value – ie, add value defined in terms of relative return?
56
CO2_with_index_2nd_amends.indd 56
23/05/2011 14:07
Introduction to Currency Risk Management and Overlay
Most existing studies of the subject fall into this category. There is, unfortunately, no conclusive answer. The results are usually heavily dependent on the sample period and the base currency of choice. The question is similar to whether going consistently long or short a currency will yield systematic returns. Q2: Does passive hedging add value in terms of risk reduction (and therefore in terms of risk-adjusted return)? The answer to this question is not necessarily data-dependent and a commonsensical answer seems to be yes. Indeed, many studies have shown that hedging is effective in reducing portfolio volatility – more so for bond and balanced portfolios than for equity portfolios. In practice, currency hedging is like shooting at a moving target, rendering the surgical removal of currency risk all but impossible. However, in the spirit of “approximately right rather than precisely wrong”, it is possible to remove a large part of currency risk in most portfolios. There are a handful of articles that highlight the limitations of currency hedging, in particular Froot (1993), which disputes the long-term (more than three years for equities and more than five years for bonds) effectiveness of hedging. Q3: Does passive hedging impair the diversification benefit of currency? A host of articles published around 1990 asserted that hedging increased the correlation between asset classes. But such claims have since been discredited. If currency movements represent white noise, uncorrelated to underlying assets, passive hedging may indeed increase the correlation between foreign and domestic asset classes. Removing such white noise will clearly have no effect on diversification. There is no definitive answer to the question whether currency moves are random noise. In practice, many people assume that they are not correlated with underlying asset returns.
57
CO2_with_index_2nd_amends.indd 57
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
Q4: (a) Does active currency management add value? (b) Are currency movements predictable? The answer to whether actively managing currency risk can add value is supported by many studies since a handful earlier surveys carried out in the early 2000s.26 Generally speaking, active currency management had added value, to the tune of 100 basis points per annum as a long-term, all-sample average. However, there has been a wide distribution of returns from different managers, due to managers’ skills, base currencies of the mandates, periods over which the mandates were traded, and benchmark hedge ratios of the mandates. There seems to be some evidence that the average returns earned from currency managers have been decreasing over the years. There are also claims that many currency managers’ returns were merely from some beta strategies. Evidence that it does comes mostly from overlay managers and investment banks. A review of evidence is included in Chapter 2, Section 2.8. There is also huge body of material on the predictability of currency movements, and this is reviewed in Chapter 4. Q5: Is the answer to the currency-hedging question different for bonds and equities? Yes and no. Currency volatility is usually much greater than bond volatility and currency exposures in global bond portfolios are often hedged as a matter of course. The risk-reduction benefit for equities is much smaller and some investors choose to leave currency unhedged. That’s the yes part. For the no part: if an investor invests in both global bond and global equity portfolios, it may make sense for them to look at their combined currency exposures rather than bonds and equities separately – we will revisit this issue in Section 3.3.2 from a practical viewpoint and again in Section 5.4, when we discuss portfolio optimisation. Q6: Can currency hedging add value by other means? We believe this to be one of the most important yet overlooked 58
CO2_with_index_2nd_amends.indd 58
23/05/2011 14:07
Introduction to Currency Risk Management and Overlay
issues in the currency-hedging debate, and we think that the disciplined approach associated with currency overlay is capable of bringing real benefit to an organisation. The manager of one of the largest pension funds in the world said, on the basis of their experience with currency overlay, that the greatest benefit of currency hedging is the clarification of responsibility for risk. To undertake currency hedging properly the first thing you need to do is to know, and continuously monitor, your investment positions and resulting currency exposures. This exercise of clarifying risk responsibility may well bring many benefits to the fund. In our experience, those funds that run currency-overlay programmes typically have clearly defined roles and responsibilities and their internal communication appears to be more efficient. It would be interesting to see whether this is reflected in the overall performance of such funds. 1.6.4 Survey of studies on currency hedging In this subsection we survey the empirical evidence on the question whether currency risk should be hedged. We focus on the most frequently quoted articles and, where possible, we have chosen studies by authors who are not involved in the overlay industry. The results are presented in Table 1.4. It is impossible to give detailed consideration to every article, as many adopt a unique approach in dealing with the subject. Some of these – for example, hedge ratio, Sharpe Ratio-based optimisation and parameter uncertainty – will be discussed later in the relevant sections of the book. It is, however, useful to note whether a study is ex ante or ex post. In an ex ante study, the optimisation is carried out using only information available up to that point in time, whereas in an ex post study, an optimisation may be run over the whole dataset without distinguishing what information is available at what time in time. The ex ante set-up resembles the real-life operation more closely and therefore is more relevant for the practical applications.
59
CO2_with_index_2nd_amends.indd 59
23/05/2011 14:07
Authors (date)/title of paper/ Underlying asset/period and frequency/base Currency Eun & Resnick (1988) Exchange rate uncertainty, forward contracts and international portfolio selection
Main conclusions
Other features
Full currency hedging can increase the gains from international equity diversification
o Covers CAD, FRF, DEM, JPY, CHF and GBP o Both ex ante and ex post o Shows exchange rate risk non-diversifiable due to high cross-correlations o Hedging results in significant performance improvement by reducing parameter uncertainty
Hedging significantly reduces currency risk without scarifying the expect return.
o Use quarterly real USD return27 o Not based optimisation o Full currency hedge reduces equity volatility by 20% to 30%; bonds 35% to 60%. o Authors estimate that transaction cost for a six-month rolling programme is 0.12% per annum
Equities 1979–1985, weekly USD Perold & Schulman (1988) The free lunch in currency hedging: Implication for investment policy and performance standards Bonds and equities and balanced portfolios 1977–1987, quarterly USD
CURRENCY OVERLAY – SECOND EDITION
60
CO2_with_index_2nd_amends.indd 60
Table 1.4 Should currency risk be hedged? A review of the empirical evidence
23/05/2011 14:07
CO2_with_index_2nd_amends.indd 61
Table 1.4 (continued) Thomas III (1989) The performance of currencyhedged foreign bonds
Hedging significantly improves the performance of foreign currency denominated government bond portfolios
o Ex post study covering DEM, JPY, GBP, FRF, CAD and NLG o Not based on optimisation o Currency risk decreases from average 15% per annum to 8% with hedging; at portfolio level, from 12% to 5% o Risk reduction is consistent in sub-periods
US-based investors benefit only marginally from investing in currency-hedged foreign bonds, and when hedging cost is considered, investing in foreign bonds is not worthwhile
o Using ex post optimisation over Sharpe Ratio with different asset classes o In unconstrained optimisation, optimiser does choose hedged foreign-currency bonds; however, the efficiency loss is very small when hedged foreign-currency bonds are excluded o Sharpe Ratio for US, unhedged and hedged foreign bonds are 0.22, 0.13 and 0.25 respectively in this study
Hedging is beneficial and a forward bias-based hedging strategy provides superior performance to fully hedged strategy
o Covers CAD, DEM, JPY, CHF and GBP o Ex post results
Bonds 1975–1988 USD
Foreign bonds in diversified portfolio: A limited advantage Bonds 1978–1987 USD Eaker and Grant (1990) Currency hedging strategies for internationally diversified equity portfolios Equities 1975–1988 USD
61
Introduction to Currency Risk Management and Overlay
Burik and Ennis (1990)
23/05/2011 14:07
Hauser and Levy (1991) Optimal forward coverage of international fixed income portfolio
Hedging is more beneficial for shorter-maturity foreign bonds
o Ex post study o Covers USD, JPY, DEM bonds o Currency risk decreases from 95% to 75% when maturity of foreign bonds goes from 0.5 to 5 years o Forward bias-based hedge shows the best performance improvement
Hedging is beneficial but hedge ratio decreases with cost of hedging. Hedge ratio also depends on portion of assets allocated internationally as well as investor’s risk appetite
o A pragmatic approach linking cost of hedging and hedge ratio o Based on the idea that investor is willing to give up 0.28 basis points for 1% reduction in portfolio volatility
For shorter horizon, hedging reduce volatility. For horizon longer than five/eight years, foreign stocks/bonds display greater return volatility when hedged than not
o Only one currency pair as looking at USD investments with GBP as base currency o Real returns instead of nominal returns o Minimum variance hedge ratio reduces from almost 100% at short horizon to an average of 35% at 5–10-year horizon
Bonds 1983–1988 USD Nesbitt (1991) Currency hedging rules for plan sponsors Equities 1973–1989 USD Froot (1993) Currency hedging over long horizons Equity and bonds 1802–1990 GBP
CURRENCY OVERLAY – SECOND EDITION
62
CO2_with_index_2nd_amends.indd 62
Table 1.4 (continued)
23/05/2011 14:07
CO2_with_index_2nd_amends.indd 63
Table 1.4 (continued) Glen and Jorion (1993) Currency hedging for international portfolios
Hedging is beneficial, though more so for world bond and balanced portfolios than for world equity portfolios
o Both ex ante and ex post o Including significance tests and out of sample performance o Optimised hedge ratio does not significantly outperform unitary or universal hedge ratio28 o Black’s universal hedge ratio does not provide statistically significant performance o Forward bias-based strategy provides significant improvement over a fully hedge strategy
The timing of country allocation shows that underlying managers do not have skills to forecast currency movements. A small amount of exploitable inefficiency presents in foreign exchange forward market
o Actual allocation and performance data for 18 managers makes this the only study we are aware of to use such data o Important implication on the need to separate currencies from underlying management o Forward inefficiency is the only alternative tested in the study
Rule-based active management significantly improves portfolio performance as measured by Sharpe Ratio
o Covers DEM, GBP, CHF, JPY and CAD o Incorporates transaction costs and carry costs o Uses bootstrap-based significant test o Sharpe Ratios: unhedged 0.29; fully hedged 0.38; active 0.75 o Uses filter rule and moving average rule
Bond, equity and balanced portfolios 1974–1990, monthly USD Investing abroad: A review of capital market integration and manager performance Bonds and equities 1988–1990, actual allocation and performance data of 18 managers USD Levich and Thomas (1993b) The merits of active currency risk management: evidence from international bond portfolios Bonds 1976–90
63
USD
Introduction to Currency Risk Management and Overlay
Halpern (1993)
23/05/2011 14:07
Eun and Resnick (1994) International diversification of investment portfolios: US and Japanese perspectives
For US-based investors hedging is beneficial both ex ante and ex post; for Japan-based investors, only ex post.
o Covering CAD, FRF, DEM, JPY, CHF, GBP and USD o Both ex ante and ex post o Hedging gain is more significant for bond and balanced portfolios but less so for equity portfolios o The gain for international diversification is greater for US based investors than for Japan-based investors29
Currency hedging is generally beneficial. However, overlay-style separate optimisation is not as efficient as full optimisation30
o Uses ex post optimisation over Sharpe Ratio with different asset classes o Author acknowledges that the ex post result is of little relevance to actual portfolio management o Optimisation results in hugely implausible hedge ratios, e.g., as high as 1450%; no position in German asset but taking 52% short position in DEM, etc. o Forward bias-based strategy adds value
Hedging reduces both portfolio risk and estimation risk considerably, more so for USD and JPY-based investors. However, for USD, DEM and JPY-based investors, hedging has been expensive due to forward bias
o Ex ante optimisation based on Sharpe Ratio, closely following Eun & Resnick (1988) but with bonds as underlying o Covering 15 different currencies, of which eight later joined EUR, four-track EUR closely (hence smaller benefit of hedging for DEM-based investors) o Alternative estimators, which are designed to deal with estimation risk, do not work as well when short-sale constraints are imposed.
Equities, Bonds, Balanced Portfolios 1978–1989 monthly USD and JPY Jorion (1994) A mean–variance analysis of currency overlays Bonds and equities 1978–1991, monthly USD Annaert (1995) Estimation risk and international bond portfolio selection Bonds 1985–1991, weekly USD, DEM and JPY
CURRENCY OVERLAY – SECOND EDITION
64
CO2_with_index_2nd_amends.indd 64
Table 1.4 (continued)
23/05/2011 14:07
CO2_with_index_2nd_amends.indd 65
Table 1.4 (continued) Braccia (1995)
Optimal hedge ratio is as high as 0.80 if cost of hedging is zero. An Analysis of Currency Overlays However, hedge ratio is much for US Pension Plans lower, averaging 0.30, when a 55 basis points cost is included Equity and bonds1980–1993 USD
No consistent pattern found to indicate that hedging can enhance Currency management: Concepts or reduce returns; currency and practices hedging, however, consistently reduces asset risks Equity and Bond
o Currency hedging reduces risks in equity/bonds investment by average 8% and 24% respectively o Both return and risk figures seem high in this study. For example, unhedged bonds risk typically over 20% and hedged bonds 13–19%
1980–1993 USD, GBP, DEM, FRF, JPY, AUD, CAD Eun and Resnick (1997) International equity investment with selective hedging strategies Equities 1978–1994
65
USD
Currency hedging marginally benefits the international equity investments. Active strategies such as forward bias-based hedging and writing options provide enhanced results
o Covers CAD, FRF, DEM, JPY, CHF, GBP and USD o Out-of-sample tests and ex ante strategies; uses simulation technique o Deals with parameter uncertainties o Overall, the option writing strategies performed the best
Introduction to Currency Risk Management and Overlay
Clarke and Kritzman (1996)
o Covers DEM and JPY o Portfolio consists of US and international equities, US and international bonds o Some curious correlation figures: for example, correlation between USD/JPY (USD/DEM) and international bonds is 0.84 (0.64). o Author explains: ‘correlations among asset classes and the currencies are low enough to provide adequate diversification without incurring the added cost associated with currency hedging’
23/05/2011 14:07
A full currency hedging strategy combined with minimum-variance optimisation on local currency excess returns outperforms other hedging/optimisation combinations
o Ex ante study, covering equity indexes from 14 countries o Definition of local currency excess return similar to that of Karnosky and Singer (1994 – see Chapter 6, Section 6.6 o Results consistent for investors based in different currencies
Passive hedging reduces portfolio performance. However, a moving average rule outperforms both hedged and unhedged portfolios
o Ex post, covering FRF, DEM, JPY, GBP and USD o Sharpe Ratio is reduced from 0.26 to 0.18 when currency is fully hedged for a portfolio with 40% of foreign asset allocation o Local currency indexes are used as the “hedge indexes” (see Chapter 6, Section 6.4)
For US-based investors, full hedging does not improve Sharpe Ratio for equities but it does for Conditional hedging and portfolio bonds. Results are different for performance different base currencies
o Ex post optimisation o A forward bias-based strategy, using both real and nominal interest rates, performs better in all cases o In bond portfolios, Sharpe Ratio uniformly improves with higher hedge ratio. In equity and balanced portfolios, it peaks before declining for higher hedge ratio
Larsen and Resnick (2000) The optimal construction of international diversified equity portfolios hedged against exchange rate uncertainty Equities 1983–1988, monthly USD, FRF, DEM, NLG and CHF Reiner (2000) Practical active currency management for global equity portfolio Equities 1972–1999, monthly USD VanderLinden, Jiang and Hu (2002)
Bonds and equities 1976–1997 23/05/2011 14:07
FRF, DEM, JPY, GBP, USD
CURRENCY OVERLAY – SECOND EDITION
66
CO2_with_index_2nd_amends.indd 66
Table 1.4 (continued)
CO2_with_index_2nd_amends.indd 67
Table 1.4 (continued) de Roon, Nijman and Werker (2003) Currency hedging for international stock portfolios: The usefulness of mean-variance analysis Equities
Hedging is utility-dependent. Passive hedging is only beneficial for investors with very high risk aversions. It is not beneficial for investors with mean-variance utility. Forward bias-based hedging leads to significant performance improvement
o Ex post optimisation using regression, covering G5 currencies o Hedge ratio highly variable from optimisation, sometimes running into thousands of percent points, either to hedge or to speculate.
USD Chang (2009)
Full Hedge reduced total portfolio risk for JPY based investors, also for Currency Hedging: A Free Lunch? investors investing in Australian, Canadian, New Zealand, and US Equities equities. In other cases, the results are mixed. But for global portfolio, 1987–2008, monthly and daily hedging in general reduced total USD, AUD, EUR, GBP, and JPY risk, by around 10% to 20%.
o Looking at MSCI hedged versus unhedged indexes. Only 100% hedge ratio was investigated. o Hedging in many cases reduced both risk and return.
Campbell, Medeiros, and Viceira 100% hedging is not a risk minimizing strategy due to (2007) correlations of equity and currency Global Currency Hedging markets. Bonds should be fully hedged. Bonds and equities
o Investors should more than fully hedge currencies negatively correlated with local equity markets, such as the USD, and less than fully hedge currencies positively correlated with local equity markets, such as the CAD. o Equity investors should take long positions in USD, EUR, and CHF as they are reserve currencies and then to appreciate during equity sell-off.
1975–2005, Quarterly
67
AUD, CAD, EUR, JPY, CHF, GBP and USD
Introduction to Currency Risk Management and Overlay
1975–1998, monthly
23/05/2011 14:07
Brandes Institute (2007) Currency Hedging Programs: A Long-Term Perspective Equity 1972–2006, monthly CAD, EUR, JPY, GBP and USD Schmittmann (2010) Currency Hedging for International Portfolios Bond and equity 1975–2009, monthly
Hedging or not hedging can have a significant short and long term impact on portfolio return. Shortterm outcomes encompassed an extremely wide range (over +/−20%). Even at 10-year horizon, the outcomes can vary up to 10% on annualised basis, often more than +/−3%.
o It is probably unwise to adjust strategy hedge ratio too frequently. Changing it after two or three years would often lead to losses – behavioural risk. o Average active return is less than 1%, very small compared to annual potential gain or losses from passive currency hedging.
Extending Froot (1993) to multicurrency portfolios. Hedging effectiveness does not decrease with longer horizon over the recent 35 years.
o Hedging is base currency dependent. o Full hedge for bonds. Equity depends on the overall correlation between the base currency and the equities. However, the correlation structure changes substantially during the crisis.
CURRENCY OVERLAY – SECOND EDITION
68
CO2_with_index_2nd_amends.indd 68
Table 1.4 (continued)
23/05/2011 14:07
Introduction to Currency Risk Management and Overlay
Our aim in including this rather long survey is to allow readers to make up their own minds on the currency-hedging debate. The evidence available so far, as reviewed in Table 1.4, is summarised in the points below: o currency hedging is generally beneficial in terms of risk reduction, more so for bonds and balanced portfolios than for pure equity portfolios; o the impact of currency hedging on returns or risk-adjusted returns (eg, Sharpe Ratio) is unclear; o hedge ratios obtained from CAPM-style optimisation using Sharpe Ratios are generally not very practical, and the optimisation method is very sensitive to the expected returns used in the optimisation process; o in ex ante studies a unitary hedge ratio (100%) works better than optimised hedge ratios – see also Endnote 15; o managers of global equity portfolios and bond portfolio may not be able to time the moves in currency market; o the impact of hedging is different for different base currencies; o forward bias-based hedging strategies generally work, as do some other quantitative strategies; o the hedge ratio may also depend on other factors, such as investors’ risk appetite and the cost of hedging. 1.6.5 Currency overlay: what it is and what it is not The concept of currency overlay means different things to different people and currency-overlay managers, in our view, have done very little to demystify it. In many sales pitches and published articles by currency-overlay managers that we have seen, currency overlay is presented as a means to earn extra returns for next to nothing and/ or to reduce portfolio volatility, again, for next to nothing. Otherwise there is very little information on how they obtain their upwardsloping cumulative return curves (also known as “equity curves”) or downward-sloping (at least to a point) risk curves. Unstated assumptions may have been made to get these numbers, which means that the conclusions must certainly be viewed with caution.
69
CO2_with_index_2nd_amends.indd 69
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
Table 1.5 Currency overlay: what it is and what it is not It is . . .
It is or does not . . .
o One of the methods to manage
o A “magic product” that will
o A useful framework for quantifying
o Necessarily view currency as
o A tool for measuring performance
o Necessarily trade currency
o An exercise to clarify risk
o Necessarily involve financial
currency risks
currency risks in a portfolio or within an investment management institution and accommodating a variety of styles to manage them
depending on a fund manager’s core competency responsibility and to improve internal communication
with certainty bring excess returns or reduce portfolio risks a separate asset class
actively
wizardry
In this book, currency overlay is defined as a process whereby currency risk is managed separately from the underlying investment. It is, therefore, a form of currency hedging. The underlying investment can be bonds, equities, real estate and so forth. These are typically, but not necessarily, denominated in a currency other than the investor’s home currency. While we defer a detailed discussion on the subject to Chapter 2, a list of what currency overlay is and is not is presented in Table 1.5. APPENDIX 1A: UNCOVERED INTEREST RATE ARBITRAGE Uncovered interest-rate arbitrage describes the relationship between the expected returns on monies denominated in different currencies. Basically, when the expected returns are measured in the same currency they should be equal. For example, if interest rates for the Japanese yen and US dollar are 0% and 2% per annum, respectively, and the current spot exchange rate is 100.00 Japanese yen per US dollar, then in one year’s time the yen is expected to appreciate against the US dollar by 2%. This way, the return on US dollars, 2%, is the same as the expected return on the yen expressed in US dollars (0% interest rate plus 2% of appreciation). What, then, would be the covered interest-rate arbitrage? In the 70
CO2_with_index_2nd_amends.indd 70
23/05/2011 14:07
Introduction to Currency Risk Management and Overlay
above example, the return on the yen is expected, not guaranteed, since we do not know what the actual USD/JPY spot exchange rate will be in one year’s time. One can, however, effect a hedging transaction, selling ¥100 million forward at a fixed price in order to remove that uncertainty. This transaction is called the covered interest-rate arbitrage. This arbitrage relationship, not the expectations of market participants, determines the forward prices for foreign exchange. This point is discussed further in Appendix 1C. The Japanese yen, with a lower interest rate than the US dollar, will be traded more expensively against the US dollar in the forward market. The yen is also said to trade at a premium, whereas the dollar is said to trade at a discount. Violation of the uncovered interest-rate arbitrage is called forward bias. This is an important concept in foreign exchange trading and is discussed in Chapter 4, Section 4.5.1. APPENDIX 1B: THE CORRELATION TRIANGLE There is a simple geometric relationship between the volatilities and correlations of three currency pairs formed from three currencies. Let us define some notation: o σAB is the volatility of the currency pair A/B; it is expressed as the annualised standard deviation of returns on currency B measured in currency A. o AρBC is the correlation between currencies B and C with respect to currency A; specifically, it is the correlation between returns on B measured in A and returns on C measured in A. Figure 1.B.1 A geometric illustration of a correlation triangle
¥ σ$¥ $
θ
$ρ£¥=cos(θ)
σ$£
¥ρ$£
σ£¥
£ρ$¥
£ 71
CO2_with_index_2nd_amends.indd 71
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
We can now draw a volatility and correlation triangle using the US dollar, sterling and Japanese yen as an example (Figure 1.B.1). In this figure, volatility is a length and correlation is (the cosine of) an angle. Given any three products it is always possible to draw this triangle. By definition, the volatility and correlation values bear exactly the relationships shown. In particular, given any three of the values depicted, the remaining three can be determined by trigonometry. The following relationship holds: Two kinds of volatility are normally observed in the foreign-exchange market: the actual volatility and the implied volatility. In our example we could use the historical data for the past three months to calculate the actual volatilities between the pairs GBP/ USD, USD/JPY and GBP/JPY. If we then replotted the triangle and made the three sides proportional to the calculated actual volatilities, we could read off the correlation between any two currency pairs directly from the triangle (though when you do this for yourself be sure to get the sign of the correlation right). The results of this exercise are the same as the correlations estimated directly from the daily changes in the exchange rate. Similarly, you can use the three-month implied volatilities for the three currency pairs and plot the triangle once more. This time, the correlation you read off the chart is the implied correlation. Loosely interpreted, the implied correlation is where the market thinks the correlation is going to be for the next three months. If we assume that the market is efficient, then that would be the best forecast for the future correlation. If the actual and implied correlations are wildly out of line with each other, some people would make certain trades to exploit this situation. APPENDIX 1C: FORECAST, FORWARD AND NO-ARBITRAGE PRICE Some people claim that the forward foreign-exchange rate, derived from the covered interest-rate arbitrage relationship discussed in Appendix 1A, represents the market consensus forecast of where the exchange rate will be in the future. The analogy would be that the forward interest rate, derived from, for example, interest-rate futures or forward-rate agreements (FRAs), represents the market72
CO2_with_index_2nd_amends.indd 72
23/05/2011 14:07
Introduction to Currency Risk Management and Overlay
consensus forecast of where the interest rate will be in the future. Whereas the latter statement is true, the same cannot be said of the forward foreign-exchange prices as it comes via a different route: the arbitrage relationship. We will spend a little more time in Chapter 7 considering why the arbitrage relationship is so fundamental to modern financial theory. For the time being, it suffices to say that whenever an arbitrage price exists it dominates the forecast price. In other words, the prevailing market price will be the arbitrage price rather than the forecast price. Anyone quoting the forecast prices will face almost certain financial ruin. How so? Because arbitrageurs can enter into an agreement with that person on one side and use the arbitrage relationship to fix the price on the other. The arbitrageurs can therefore lock in risk-free profit – well, market risk-free at least because the person can go bankrupt and the arbitrageurs face credit risk. Here is an example: Example 1.C.1 o Spot USD/JPY rate 100.00. o One-year interest rates for US dollar and Japanese yen are 2% and 0%, respectively. o Dealer X is quoting one-year forward USD/JPY at 98.10. Note that the no-arbitrage forward price is 100 x (1 + 0%)/(1+ 2%) = 98.04.31 An arbitrageur can contract with X to sell US$1 million and buy ¥98.10 million at 98.10 in one year’s time. The arbitrageur can then borrow ¥98.04 million and sell them on the spot market at 100.00, receiving US$0.9804 million. Let us suppose that they deposit the proceeds at the 2% US dollar interest rate. In one year’s time the US dollar deposit matures to US$1 million. The arbitrageur delivers the US$1 million to X and receives ¥98.10 million. They then use this to repay the Japanese yen loan of ¥98.04 million, making a net profit of ¥0.06 million. Note that this is irrespective of what the spot USD/JPY rate might be in one year’s time, so it is free from market risk. And the arbitrageur could do any amount to multiply the profit. 73
CO2_with_index_2nd_amends.indd 73
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
What might happen to X? If they decide to hedge their spot currency risk, they will do exactly the opposite of what the arbitrageur has done and end up losing ¥0.06 million for certain. If they leave their position open, they would be exposed to any changes in the USD/JPY spot exchange rate, as they will have to sell yen and buy dollars in the spot market to settle their contract with the arbitrageur. If we assume that, almost surely, they do not get their prediction right all the time, they will almost certainly face financial ruin.32
For more information on the foreign-exchange market structure and its implications, see Lyons (2002). Also refer to Chapter 4, Section 4.6.6, for market structure and exchange-rate forecast. 2 All the BIS survey figures refer to the turnovers, not outstanding positions. The outstanding positions for interest-rate derivatives are much bigger than those for FX derivatives, because the maturities for interest-rate derivatives are usually much longer than those for the FX derivatives. 3 A clear expression of a foreign-exchange unit is currency A per currency B, where the former is the numerator currency (a.k.a. counter currency or foreign currency) and the latter is the denominator (a.k.a. base currency or domestic currency). A commonly used short form is currency B/currency A, and this convention is followed throughout the book. 4 See Appendix 1A of this chapter for an illustration of how to calculate the forward exchange rate from the spot rate and domestic/foreign interest rates. Also, in our examples we use simple compounding for simplicity. 5 Here we assume that there is no uncertainty regarding the outcome of the investment. Obviously, in practice, both the yield and the capital may well be different after three months. This is dealt with in the second example. 6 Capital usage for currency trading is usually not as easily defined as, say, that for cash investments in bonds and equities. Some argue that currency trading requires zero capital as it is conducted entirely on credit (they might not be of the same opinion if they actually tried to obtain credit lines from, say, a bank to trade currencies). Clearly, using the notional amount of currency trades as capital usage may not be appropriate either, as relatively few currency trades are actually funded this way. The answer lies somewhere between these two polar cases. Many funds employ a VaR-based approach similar to those used by banks to compute their capital requirements. This is a general problem for any fund using any form of derivative instrument and therefore is not dealt with in this book. 7 See Appendix 1A. 8 The β of an instrument is its standardised covariance with its class of instruments as a whole. Thus the β of a stock is the extent to which it follows movements in the overall market (after Chase/Risk Publications, 1996). 9 See, for example, Prajogi, Muralidhar and van der Wouden (2000) and Muralidhar and van der Wouden (2000). 10 MSCI EAFE Index: a commonly used international equity performance index originally calculated by Morgan Stanley Capital International Inc but later MSCI was spun off. EAFE stands for Europe, Australia and Far East. 11 Risk may include other aspects of the return distribution which the variance measure adopted by Froot (1993) may not be sufficient to capture. For example, the risk can be defined as extreme losses. 1
74
CO2_with_index_2nd_amends.indd 74
23/05/2011 14:07
Introduction to Currency Risk Management and Overlay
12 A functional currency is the currency in which a corporation chooses to report its results. A corporation may have several functional currencies, and the functional currency can be different from that of the country where its shares are listed. 13 The delta of an option describes its premium’s sensitivity to changes in the price of the underlying (Chase/Risk Publications 1996). 14 See, for example, Knez, Litterman and Scheinkman (1994) and Litterman and Scheinkman (1991). 15 We define currency hedging as using financial means to neutralise the impact of currency fluctuations on returns, denominated in a base currency, from investment denominated in, or impacted by, foreign currencies. Currency overlay is a method of implementing hedging. 16 77% is Black’s universal hedge ratio; see Chapter 5, Section 5.5 17 See Endnote 13 of Chapter 5 on Siegel’s paradox for an illustration of percentage gains. Appendix A of Chapter 4 contains an example where both central bankers and speculators make money out of intervention but over different horizons. 18 Commonly used factors in the academic literature include excess market returns, market capitalisation relative to GDP, balance of trade relative to GDP, inflation, changes in industrial production and credit spread. In a global setting, the factors can be excess return on a world market index, world dividend yield, world inflation, changes in world industrial production, slope of the US interest-rate curve, oil price, US corporate bond–treasury spread, etc. 19 Some commentators therefore label currency as an exposure rather than an asset. We do not make that distinction in this book. 20 See Chapter 5, Section 5.2.2.3 for a discussion of currency risk premium and references. 21 See the “peso effect” in Chapter 4, Section 4.5.1. 22 See, for example, Banerjee (1992) and Shleifer and Summers (1990). 23 For references see Baillie (1996), Andersen et al (2001), Ederington and Lee (1993) and Granger and Ding (1996). 24 Baele, Pungulescu and Horst (2007) and Sercu and Vanpée (2007). 25 According to InterSec’s estimate, the cross-border pension fund assets had increased from around US$1 trillion in 1995 to over US$3.5 trillion in 2005. 26 Hersey and Minnick (2000) and Baldridge, Meath and Myers (2000). 27 Perold and Schulman (1988) contend that quarterly data is better suited for this kind of analysis, as risk measures calculated over shorter time windows (eg, monthly or weekly) tend to underestimate the risk, because they fail to account for the changes in mean. 28 In Glen and Jorion’s setting a passive foreign-exchange hedge ratio is chosen and implemented. Then additional currency positions are taken on the basis of the result of CAPM optimisation. They show that the optimisation procedure does not provide additional value in the out-of-sample tests. The result corroborates practical observations that most currency managers do not use CAPM-type optimisation to determine currency weights. 29 During the sample period, the US dollar appreciated against the Canadian dollar, French franc and sterling and depreciated against the Deutschmark, Japanese yen and Swiss franc. The yen appreciated against all other currencies. It is puzzling why under the circumstances hedging did not improve the portfolio performance. Closer inspection of the data shows that the yen-denominated domestic bond portfolio had the highest Sharpe Ratio, which proved difficult to beat. 30 See Chapter 5, Section 5.4 on optimisation styles. 31 We use a straightforward interest-rate calculation here for simplicity. 32 In fact, market forces will, in all probability, make sure that he does not survive all that long.
75
CO2_with_index_2nd_amends.indd 75
23/05/2011 14:07
CO2_with_index_2nd_amends.indd 76
23/05/2011 14:07
2
Currency-Overlay Management
2.1 INTRODUCTION This chapter examines the concept and practice of currency overlay. After a brief historical summary, we review the different approaches investors take in dealing with currency risks in global investment portfolios, their respective pros and cons, and introduce the currency-overlay approach. We then give a definition of currency overlay in a very broad context and discuss the difference between currency overlay and currency hedging, and different styles of currency overlay. We then go into details on the key theme of the book: the need for risk awareness and a disciplined approach to currencyrisk management. In our opinion, the most vital question in the currency-risk-management debate in a portfolio management setting is not whether we should hedge risks, but rather whether currency risk is clearly defined and quantified. The answer to the question whether to hedge or not will be the by-product of a clear definition of currency risk and the existing risk/return policy of the portfolio. We present generic illustrations to show how a currency-overlay programme may function in practice. Section 2.6 reviews some of the main “styles” used for active currency-overlay programmes and currency-pure-alpha programmes, commenting on their recent developments. We then provide an overview on the main parties involved in the currency-overlay business: overlay managers, endusers, counterparty banks, custodians, etc. In Section 2.8, we survey old and more recent evidence on whether currency-overlay manag77
CO2_with_index_2nd_amends.indd 77
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
ers made money for their investors. Section 2.9 reviews practical aspects of outsourcing currency management: legal documentation, costs and how to evaluate the services on offer. The topic of transaction costs within this section is particularly important given the surprisingly opaque nature of the foreign-exchange market. 2.2 ORIGINS AND DEVELOPMENT OF CURRENCY OVERLAY The concept of currency overlay first surfaced in the late 1980s, when specialists began to provide, initially in the US, disciplined currencymanagement services to the pension fund sponsors as well as fund managers with international investment portfolios. The big US dollar cycle during the 1980s provided the backdrop for this development: in the first half of the decade, many US dollar-based fund managers started to invest globally but their returns were severely crimped by the rising US dollar. Some of them elected to put full currency hedges on their portfolios. The hedging was conducted by either the underlying managers or the custodian banks, or by commercial banks/investment banks offering hedging service. It was usually done on an ad hoc basis: hedging decisions were taken on individual batches of investment, based on asset classes, nature of the mandates (internally managed or externally, for example), or just organisational timeline (for example, an investor may have recently increased their allocation to global bonds; they were concerned about the increasing level of foreign-currency exposures and therefore decided to hedge the new allocation in global bonds). The patchwork of hedging grew in complexity as the size and complexity of the underlying portfolios grew, and some investors, investment managers and service providers began to ask whether this could be done more systematically and efficiently. It is natural that the concept and practicality of currency overlay was born then. The first mandates were awarded in the UK, Japan, the US and continental Europe in 1985, 1987, 1988 and 1992, respectively, according to Folkerts-Landau (2002). Looking back, we can see that currency overlay could have been called “systematic currency hedging”, or, plainly, just “currency hedging”, but calling a spade a spade is not how the financial industry does things. It is now difficult to attribute the invention of currency overlay to a particular person or organisation – the author has come across three people so far known as the “grandfather of currency overlay”. 78
CO2_with_index_2nd_amends.indd 78
23/05/2011 14:07
Currency-Overlay Management
Those US-based investors hurrying into the latest investment trend, “currency overlay”, which was really a “passive currency overlay” with a full hedge ratio, towards the tail end of the US dollar rally were hurt on the reversal of the trend during the second half of the decade. But the argument was that hedging was not meant to make money anyway, that what you had lost on the hedge during the period of weak US dollar is merely a mirror image of what you had gained on the underlying global investment due to currency moves, revalued in US dollar terms. There was a snag, though: what you gained was an accounting entry, not realised until you liquidate the foreign investments and bring the money back; what you lost was to be paid in cash. This is known as the “cashflow problem of hedging”; some commentators use it to argue against currency-hedging practice, saying that currency hedging only shifts the currency volatility suffered by the underlying portfolio into cashflow volatility. We will deal with this issue in Chapter 3, Section 3.3.7. The US dollar weakness was to last for a while, investors writing cash cheques month after month started either to close their currency-hedging programmes or reduce their hedge ratios. The concept and practice of “active currency overlay” was created: it would be ideal if the hedge ratio is adjusted in anticipation of the currency moves, rather than forced to be adjusted because of painful losses. With active currency overlay came currency specialists in the real sense: you need people, or software programs written by people, to pre-empt the currency moves. The currency specialists are supposed to watch the currency market like hawks, trying to make money based on the existing currency exposures in global investment portfolios. The marketing materials from these specialists might have said, “Turn foreign-exchange risk into opportunity”; “Let your existing currency exposures work for you”; “A freelunch: How to reduce risk and get extra returns at the same time”; “Express your global macro views efficiently via currencies.” In the 1990s, the method of currency overlay had gained acceptance outside the US, by pension funds, insurance companies and mutual funds that managed the assets from the pension funds and insurance companies. What was a bad market environment for investors based in one currency (eg, a US dollar depreciation period 79
CO2_with_index_2nd_amends.indd 79
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
for US-dollar-based investors) can only be good for investors based in other currencies. Countries with established and sophisticated pension-fund industries tended to use currency overlay more extensively: the US, Canada, the Netherlands, Sweden, Switzerland, Belgium, Australia and Japan. The specialist currency-overlay industry grew in size and depth. The top players include both the integrated houses, where currency overlay is offered as an additional service to the management of bond and equity portfolios, and the pure currency specialists, where only a currency overlay service, in its many forms, is offered. The late 1990s saw the emergence of the hedge fund industry as a viable asset class for institutional investors, some currency-overlay specialists started offering standalone currency products, which at the time were usually termed currency hedge funds but later categorised as “currency pure alpha”. These products are an extension to the active currency-overlay programmes, but with an important difference: currency-pure-alpha products are not directly linked to the existing currency exposures in an underlying portfolio, whereas currency-overlay programmes are. The overlay industry argued that, by severing the link between the underlying exposures and the maximum positions in any given currency, thus freeing up the currency managers, they could generate more alpha by deploying risk capital on the most promising currency opportunities rather than on the currencies where the underlying portfolio has the largest exposures. This trend, known as “alpha–beta separation” and the topic of Chapter 7, gained momentum through out late 1990s up to the new millennium. Many currency-overlay managers started offering alpha products, drawing experience and track record from their active overlay programmes. Some branched out, offering an overlay service including currencies, bond and equity indexes, better known as “global tactical asset allocation” (GTAA). The lean years for currency trading over 2006 to 2008, when volatility was low but heading lower, just before the onset of the global financial crisis, did not provide fresh impetus for investors to go into currency pure alpha. And the crisis itself put the trend into reversal. The overlay industry is still growing, but at a slower rate compared with late 1990s and early 2000s, especially the outsourced currency-overlay programmes. In recent years more notable areas 80
CO2_with_index_2nd_amends.indd 80
23/05/2011 14:07
Currency-Overlay Management
of growth have been (1) from the institutional investors in emerging markets and (2) from global institutional investors that have set up internal currency-overlay programmes – many such institutions have now put in place specific foreign-exchange policies and processes that are in effect currency-overlay programmes in all but name. The latter development is not really captured in the assetsunder-management (AUM) figures from the overlay industry but provides strong supporting evidence for one main reason for currency overlay: risk responsibility, the topic of Section 2.4.4. Auxiliary business areas have been catching up too. Custodian banks now offer consolidated currency-exposure reports, ready for currency-overlay programmes, as part of their service; some also offer combined custodian and passive overlay packages to institutional investors, for a reduced fee. Investment banks offer primebrokerage services to facilitate the settlement process for currency overlays and provide additional mid- and back-office support. Some investment banks also offer a passive overlay transaction service based on fixing exchange rates and this service can be used by the investors directly or through an overlay manager indirectly. Asset consultants have collected performance data on currency managers that can be used by clients in their search for suitable currency-overlay managers. Index providers now routinely offer currency hedge indexes, customised for specific base currencies and hedge ratios and other hedging particulars. Even the exchanges are now considering offering clearing service to currency-overlay managers. We will return to these topics later. 2.3 DIFFERENT APPROACHES TO MANAGING CURRENCY RISKS Before we give a formal definition of currency overlay and discuss its many variants, let us make a little detour to go through the typical choices made by institutional investors on currency questions – given the diverging views on how to manage currency risks, it should come as no surprise that there are many different approaches. In this section we summarise the pros and cons of the different management approaches we have seen in practice. 2.3.1 A do-nothing approach This is still one of the most common approaches adopted by inves81
CO2_with_index_2nd_amends.indd 81
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
tors in dealing with currencies, and the rationales cited include the zero-sum game, random walk and long-term wash characteristics of currency movements. However, a fine yet fundamental distinction has to be made here. o Some investors/investment managers have decided to leave their currency exposures unmanaged after detailed reviews of the fund’s investment styles, its objectives and constraints in addition to empirical and theoretical evidence on the pros and cons of currency management. We think this is a valid approach since, as mentioned earlier, evidence on the benefit of currency management, active or passive, is mixed and it is a judgement call for a fund to weigh the cost and benefit of hedging. However, the important thing is that currency risk is recognised and measured and hopefully communicated to the stakeholders in the fund. Moreover, all investment decision makers are measured independently from currency moves. Namely, the performance-measurement system should be such that they can neither claim credit on favourable currency moves nor blame poor performance on the unfavourable ones. In real life such a clean-cut system may be difficult or impossible to design or implement, but the influence of currencies on the measurement of investment decisions should be minimised through the use of suitable benchmarks – a topic for Chapter 6. o Other investors/investment managers simply decide to leave their currency risk exposure unmanaged without going through a process to define, quantify the level of risk involved, nor to design and implement a performance-measurement system. We believe that this is a head-in-the-sand approach and, although in the long term currency impact may indeed cancel it out due to pure luck, this approach leaves an important source of risk unaccounted for. The approach can also be manipulated by some investment managers to hide the consequences of poor investment decisions and incompetence. We have seen many institutions switch from this type of do-nothing (which we label “uninformed” do-nothing) approach to the first type (“informed” do-nothing) over recent years and the trend continues. 2.3.2 Active versus passive versus dynamic hedging With a passive management style we typically refer to a foreign82
CO2_with_index_2nd_amends.indd 82
23/05/2011 14:07
Currency-Overlay Management
exchange benchmark with a fixed hedge ratio; foreign-exchange managers will automatically execute hedging orders to that fixed hedge ratio and they have no flexibility in terms of deviating from the benchmark. An active management style is generally employed in addition to a passive one. In this style the foreign-exchange manager can deviate from the benchmark hedge ratio by underhedging or overhedging and/or take positions in currency pairs where the fund does not have an underlying position. There are also various foreign-exchange management styles that fall between these two. For example, in static option replication1 the manager calculates a delta using an option-pricing model and adjusts the hedging automatically. This is known as a “dynamic hedging” style. It is passive in the sense that the foreign-exchange manager does not deviate from the benchmark. But it is also active because the benchmark hedge ratio changes with change in the spot exchange rate and other factors. In the next three subsections we briefly summarise the advantages and disadvantages of these management styles. 2.3.2.1 A passive hedging approach The pros: o A passive style is easier to implement and monitor; o If the fund is not very big, there is no need to have foreign-exchange experts, because bond portfolio managers can manage the currency risks; o It is more intuitive to communicate to investors/sponsors/fiduciaries (people are more likely to accept the zero-sum-game argument that foreign-exchange returns average out over time than go through technical details on why foreign exchange is special and how it can add return); o It has a relatively low direct cost. The cons: o A passive style is not able to capture additional returns; 83
CO2_with_index_2nd_amends.indd 83
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
o The manager cannot react to short-term opportunities, which are characteristic of the currency market (for example, the long-run correlation between currencies and equities is typically close to zero, but in the short run it may be positive or negative for a considerable time; a passive management style will not be able to respond to this and may actually increase the overall risk of the fund); o It may lead to cashflow problems, as losses on the hedging contracts need to be settled in cash, and a passive management style cannot adjust to adverse cashflow situations. We will see in Chapter 7, Section 7.3, that the decision on passive hedge ratio may well have the biggest consequence in terms of the risk/return profile of a fund. This is true for active and dynamic approaches too, since you need to define passive hedge ratio for those two approaches as well. 2.3.2.2 An active hedging approach The pros: o Active management gives scope for yield enhancement. Various short-term opportunities such as trends, forward premiums and technical predictions can be exploited; o In-house specialists’ performance can be benchmarked to publicly available currency benchmarks; o Part of the risk exposure can be outsourced; o Risk reduction (without giving up expected returns) can be pursued as a separate objective; o Cashflow can be a decisional variable; o Last but not least, a fund employing an active approach can easily adapt to the risk-budgeting process. Namely, it allocates its assets according to the active level of risk it runs in different risk classes. The cons: o An active style is slightly more difficult to implement and monitor as there will be additional “active positions”; o One needs currency specialists; 84
CO2_with_index_2nd_amends.indd 84
23/05/2011 14:07
Currency-Overlay Management
o Investors/sponsors/fiduciaries have to be convinced that active management does not equate to taking wild bets on currencies; o There is also a need for relatively sophisticated risk-measurement and -control systems (though this may be an advantage given the increasingly strict reporting requirements for pension funds around the world). 2.3.2.3 A dynamic hedging approach The dynamic approach is linked to option-pricing theory. An option can be viewed as a hedge plus an insurance contract in case the exchange rate goes against the hedge, all for an upfront premium. The dynamic approach attempts to replicate the payoff profile of any option but actively adjusts the hedge ratio according to some specific rules that will be discussed in Chapter 8. The aim is to own an option without paying the option premium – at least, not all of it. A similar approach exists for equity portfolios, and it is sometimes called portfolio insurance, a technique blamed by some for the stock market crash in 1987. In a nutshell, a dynamic approach tries to beat the market – enjoying the benefit of an insurance contract without paying the market price for the contract. We will review the relevant empirical evidence for its effectiveness, again in Chapter 8. The pros: o As you hedge with an insurance contract, there is no worry whether a currency goes up or down; o The cashflow problem associated with currency hedging can be mitigated (relative to a passive hedge, you will have a smaller hedge contract to settle when the exchange rate moves unfavourably); o There is no need to pay a large upfront option premium; o Sometimes a currency options market does not exist or is not liquid enough to facilitate an option-based hedging transaction, in which case the “DIY” hedging provides a viable alternative; o There is no need to hire a full team of researchers/analysts/decision makers, as the dynamic approach is still relatively straightforward. The cons: o Beating the currency options market may not be easy (and there 85
CO2_with_index_2nd_amends.indd 85
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
is no conclusive evidence that it can be beaten); o Now that options markets are becoming increasingly commoditised, success in the past is no guarantee of success in the future; o A dynamic approach is not as easy to communicate to investors; o How frequently you undertake dynamic hedging is largely an empirical optimisation problem and is based on the premise that history will repeat itself. 2.3.3 The currency-pure-alpha approach Investors adopting a currency-pure-alpha approach treat currency as a separate asset class – a tactical or alpha asset class following our discussion in Chapter 1. In a currency-pure-alpha programme, the investor or the investment manager would take currency long/ short positions irrespective of the underlying currency exposures. In fact, whether the underlying exposures exist does not matter. The currency positions are taken purely for excess returns. The positions can be in either mature currencies or emerging-market currencies, and can be (and usually are) leveraged. In comparison, hedging positions are normally restricted to dealing in predefined currencies in which underlying investments are made, and are usually not allowed to be leveraged. It is worth noting that an investor can take one of the aforementioned hedging approaches in addition to a currency-pure-alpha approach. In fact, many investors do use passive currency hedging for risk control and a currency-pure-alpha approach for return enhancement. 2.4 DEFINITIONS AND MERITS OF CURRENCY OVERLAY There are many different interpretations of the currency-overlay concept. At one end of the spectrum, currency overlay means that currency risk is hedged. At the other end it means that currency is viewed as a separate asset class and positions in currencies are taken for speculative gain. And in between flourish various hybrids. Some view a passive currency-hedging programme with a fixed hedge ratio as the most basic form of currency overlay. In this type of programme, existing hedges are periodically adjusted according to changes in the value of the underlying assets and rolled forward. These hedges will (partially) neutralise the fluctuation in the value of the foreign-currency investment caused by exchange-rate 86
CO2_with_index_2nd_amends.indd 86
23/05/2011 14:07
Currency-Overlay Management
movements. Generally, the hedge ratio is set beforehand, possibly through an optimisation process, and it is fixed irrespective of overall trends in the currency market. Risk control is the main objective of such a passive hedging programme. In many sales pitches by overlay managers, currency overlay is described as a way of actively adjusting hedge ratios in anticipation of exchange-rate moves: keeping currencies fully hedged when the exchange rate is expected to move unfavourably and fully unhedged otherwise. The rationale behind this is that currency moves are governed by a different set of variables from traditional asset classes. Specialists can use their skills, based on an intimate knowledge of how the currency market functions or on superior trading models, to predict currency moves. Seeking extra return while controlling the currency risk are the two-fanged objectives of such an active hedging programme. Historically, the concept of currency overlay started as a passive tool to control the portfolio risk; through late 1990s and early 2000s, the concept of currency overlay was increasingly linked to the twin objectives of yield enhancement and risk control. During the 2000s, more and more investors followed the “alpha–beta separation” method and started using a combination of a passive hedging programme for risk control and an unrestricted active currency programme for return enhancement – a topic for Chapter 6. The motivations for an investor to use currency overlay are clear: to reduce currency risk and/or to enhance returns. Currency-overlay managers usually use the following statements to make case for (active) currency-hedging programmes. o Most state, implicitly or explicitly, that the currency market is inefficient and hence provides arbitrage opportunities. The inefficiencies manifest in many different ways: trending, forward bias, mean-reversion, range-boundedness, responsiveness to various signals, etc. o Some claim that, by actively managing currency risks, they can reduce the impact of a negative cashflow related to a passive hedging programme. o Some say that by concentrating on the major currency exposures
87
CO2_with_index_2nd_amends.indd 87
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
in a portfolio they can reduce transaction costs and exploit the diversification effect between currencies. o Others believe that they can reduce the risk of a currency-induced valuation disaster by adopting a dynamic approach.
Figure 2.1 Payoff profile of a risk-management-motivated currency-overlay programme
2. Currency returns from managed portfolio 1. Currency returns from un-managed portfolio
3. Risk of large unfavorable currency impact 4. Expected return from unhedged currency exposure
Currency Return Distribution
The approach is illustrated in Figure 2.1, where the dashed curve shows the shift in the distribution of currency returns that is the aim of an overlay programme. The symmetrical, solid line follows a normal distribution, representing the random-walk view of currency returns. The dashed curve is skewed to the right, indicating that the likelihood of large losses is reduced, whereas the likelihood of large profits is increased. There may also be an increase in the likelihood of small losses. 2.4.1 A working definition Overlay (noun): a covering; anything laid over to alter the appearance. We now set out our own definition of currency overlay: Generally speaking, if the decision on currencies is taken separately from the other asset allocation and investment decisions, and if this
88
CO2_with_index_2nd_amends.indd 88
23/05/2011 14:07
Currency-Overlay Management
currency decision is executed by internal or external specialist managers with predefined risk/return objectives and other operational parameters, then this process is called currency overlay and the specialist managers are known as overlay managers.
Some refer to the currency decision as the “investment decision”, while others prefer to call it the “hedging decision”, and in practice the “specialist managers” are sometimes fixed-income portfolio managers. It is useful to note that the concept and technique of overlay is not confined to currencies: for example, a fund might run a derivatives overlay on top of its normal asset allocation process. The derivatives-overlay manager can use, say, futures contracts and/or bond–equity total-return swaps to alter the bond–equity split of their investment portfolio – this method is widely used by managers of global tactical asset allocation (GTAA), where the decisions on currencies may also be included.2 Another form of overlay, known as “equitising” or “bondising” overlay, is sometimes put together with currency overlay and we will discuss those when we deal with the cashflow issues of currency overlay in Chapter 3, Section 3.3.7. The key ingredient in the above definition is that the decision process for currencies is separated from other investment decisions; currency decisions may, however, be formalised in the same fashion as other types of strategic and/or tactical investment decisions taken by the investor. At its core, currency overlay is about separation. Many investment decisions taken by an investor or a manager will have implications on the currency exposures of a portfolio: changing the split between bonds and equities, adjusting the country allocation, favouring one security over another, and so on. The idea of currency overlay is to separate the currency elements from all these decisions, aggregating them, and deal with the net currency exposure of the portfolio on a centralised basis. In Chapter 5, Section 5.4, we will see that taking a separate decision does not equate to taking the decision separately. Currency decisions can be made at the same time as the decisions on other asset classes. Or they can be made after those decisions and when implicit currency exposures are known. The overlay programme will have a predefined risk/return objective, risk control parameters and performance measurement procedure.
89
CO2_with_index_2nd_amends.indd 89
23/05/2011 14:07
CURRENCY OVERLAY – SECOND EDITION
2.4.2 The difference between currency overlay and currency hedging Currency hedging refers to a situation where there is an existing exposure in one currency (foreign currency) and the value of this exposure when expressed in another currency (base currency) can fluctuate due to exchange-rate moves, and an opposite currency position (“hedging position”) is taken, usually with currency derivative contracts. The value of the exposure combined with the hedging position will then fluctuate less when expressed in the base currency. When the size of the hedging position equals the size of the exposure (namely, a full hedge), the net position will be neutral to any future exchange-rate moves. Currency hedging is a method used by both multinational corporations in dealing with their multicurrency cashflows and balance sheets, and institutional investors in dealing with their global investments. In the case of the latter, currency hedging can be applied to a single investment of a foreign security, combined investments in a single currency, investment portfolios in an asset class and/or those managed by a single manager, or to any combination of the above, or to a total portfolio of an investor. Currency overlay usually refers to a situation where an institutional investor and/or investment manager deals with currency risks in their global investment portfolios separately from the investment decisions taken on other asset classes. Yes, a currency overlay in the end will need to conduct currency-hedging transactions, fully or partially, but currency overlay is more than just mechanical hedging. A currency overlay is typically applied to a portfolio of underlying investments rather than single positions, whereas currency hedging can be and usually is applied to a single position. How a currency overlay is carried out is usually part of the investment policy, in which how currency risks are separated from other types of risk is clearly defined. In contrast, currency hedging is usually explained in the execution policy; the decision to run a currency overlay is a strategic one, made with a mediumto-long-term horizon in mind, while currency hedging can be ad hoc, tactical and short-term. Some people use the term “currency hedging” to mean “currency overlay” in an attempt to avoid jargon. This is not wrong, as long as the context is explained. 90
CO2_with_index_2nd_amends.indd 90
23/05/2011 14:07
Currency-Overlay Management
2.4.3 Different types of currency overlay Currency overlay usually comes in two types: o Passive currency overlay: The combined currency exposures of the underlying portfolio(s) are hedged back to the base currency using a fixed hedge ratio (“benchmark hedge ratio”), chosen by the investor. The currency-overlay manager needs only to make sure that the actual hedge ratio is at or close to the benchmark hedge ratio; o Active currency overlay: The investor will still specify a benchmark hedge ratio; however, the currency-overlay manager can decide to overhedge underhedge relative to the benchmark hedge ratio in anticipation of future moves in currencies. Usually, if the base currency is forecast to rise over the next trading session(s), then the currency-overlay manager will overhedge; and vice versa. A few other “types” of currency overlay are sometimes referred to: o Enhanced currency overlay: For active currency overlay, the range of actual hedge ratio is usually between 0% and 100%. In other words, the overlay manager cannot hedge more than the total value of the underlying exposures (hedge ratio >100%), nor can he increase the currency exposure by buying rather than selling foreign currencies (hedge ratio rUSD?
S&P
Nikkei
Nikkei
Currency decision
$eUSD/ JPY > 0?
FX Position*
Hedge
Unhedge
11%
12%
10%
Portfolio return
* Investor will invest fully in S&P and will also take a currency position to buy Japanese yen and sell US dollars despite the fact that there is no underlying position in the yen.
Assuming a 50% currency-hedged benchmark and using the parameter values in Example 1 from Table 6.2, we present the following example on attribution. Table 6.5(a) contains the allocation and return information needed for the exercise. Column (4) was calculated using the 50% hedge ratio from column (1). Columns (3) and (9) are the underlying and currency deviations from the respective benchmarks. Note that benchmark returns for underlying and currency in the bottom half of Table 6.5(b) are calculated using (1) and (4) in Table 6.5(a), respectively.
309
CO2_with_index_2nd_amends.indd 309
23/05/2011 14:09
Underlying security
Currency Selection
Benchmark
Active
Difference
Currency benchmark
Active weighting
Cash
Hedge
Currency weighting
Difference
(1)
(2)
(3)=(2)−(1)
(4)
(5)=(2)
(6)
(7)
(8)=(5)+(6)+(7)
(9)=(8)−(4)
Nikkei
45.00%
70.00%
25.00%
22.50%
70.00%
0.00%
−60.00%
10.00%
−12.5%
S&P
50.00%
20.00%
−30.00%
77.50%
20.00%
10.00%
60.00%
90.00%
12.5%
5.00%
10.00%
5.00%
N/A
N/A
N/A
N/A
N/A
N/A
100.00%
100.00%
0.00%
100.00%
90.00%
10.00%
0.00%
100.00%
0.00%
USD cash
Table 6.5(b)6 Currency-surprise attribution example: return assumptions Unhedged (loc ccy)
Nikkei S&P
Risk-free (USD)
Forward premium
Spot return
Hedged (USD)
Ccy Surprise
(a)
(b)
(c)
(d)
(e)=(a)+(c)
(f)=(d)−(c)
14.00%
9.00%
−7.00%
−5.00%
7.00%
2.00%
8.00%
2.00%
0.00%
0.00%
8.00%
0.00%
CURRENCY OVERLAY – SECOND EDITION
310
CO2_with_index_2nd_amends.indd 310
Table 6.5(a) Currency-surprise attribution example: benchmark and actual portfolio allocation assumptions
23/05/2011 14:09
CO2_with_index_2nd_amends.indd 311
Table 6.5(b) (continued) USD cash
2.00%
2.00%
0.00%
0.00%
2.00%
0.00%
−2.25%
7.25%
0.90%
Benchmark returns (computed)* Underlying
10.40%
Currency
0.45%
* Underlying and currency benchmark returns calculated using (1) and (4) in Table 6.5(a), respectively. Benchmark Design and Performance Measurement
311
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
From the data in the tables it is easy to compute the benchmark total return as 10.40% + (−2.25%) − 0.45% = 7.70% and the actual portfolio total return as 70% x 14% + 20% x 8% + 10% x 2% + 70% x (−5%) + 20% x 0% + 10% x 0% + ((−60%) x 2% + 60% x 0%) = 6.90% How do we attribute the −80bp difference between the actual portfolio return and the benchmark return? This is shown in Table 6.6.7 Table 6.6 Return attribution in currency-surprise example Underlying Nikkei
Currency
Total
−0.06%
−0.14%
=(70%−45%)*(7%−7.25%)
=(10%−22.5%)*(2%−0.90%)
−0.23%
−0.11%
=(20%−50%)*(8%−7.25%)
=(90%−77.50%)*(0%−0.90%)
−0.26%
0.00%
−0.26%
−0.25%
−0.80%
S&P
USD Cash
−0.20%
−0.34%
=(10%−5%)*(2%−7.25%)
Total
−0.55%
6.6 THE EXCESS-RETURN APPROACH Karnosky and Singer (1994) proposed an alternative methodology that is consistent with the currency surprise and forward premium framework, although their framework is more general than in the currency-surprise approach. The main advantage is that the new approach can easily accommodate the cross-hedging positions, whereas the currency-surprise approach requires a cumbersome calculation for bilaterally hedged returns. Central to their approach, 312
CO2_with_index_2nd_amends.indd 312
23/05/2011 14:09
Benchmark Design and Performance Measurement
they view an investment in a given currency as a combined holding of cash and an asset. The expected return on the cash part is the local risk-free interest rate. The expected return on the asset is the local currency return above the local risk-free interest rate, ie, the excess return. Under this definition, the currency-hedging decision has a very straightforward interpretation. Portfolio managers generally hold cash as part of strategic allocation and/or operational purposes (see Chapter 3, Section 3.3.7. The cash may be in a mix of currencies and the return on the cash balances should be close to the risk-free interest rates in the respective currencies. It is easy to work out that if the cash holdings in nondomestic currencies are hedged, their returns will be almost exactly the same as the domestic risk-free interest rate. Namely, the cash return in a foreign currency adjusted by the forward return (premium or discount) equals the cash return in the domestic currency – a relationship that is illustrated in Appendix 2A of Chapter 2. Table 6.7 Optimal portfolios using the excess-return approach Variable
Symbol
Example 1
Example 2
Example 3
S&P
rUSD−dUSD
6%
6%
4%
Nikkei
rJPY−dJPY
5%
10%
5%
USD
dUSD
2%
2%
4%
JPY
dJPY+eUSD/JPY
5%
0%
5%
Asset decision
(rUSD−dUSD) > (rJPY−dJPY)?
S&P
Nikkei
Nikkei
Currency decision
dUSD > JPY (dJPY+eUSD/JPY)?
USD
JPY
12%
10%
Security excess return
Cash return in USD
Portfolio return
11%
313
CO2_with_index_2nd_amends.indd 313
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
There is, however, another form of cash. When a non-cash investment position is taken – be it in equities or bonds, domestic or foreign8 – the implicit aim is to seek a return above the risk-free interest rate in that currency, the latter being the return on an equivalent cash position in the currency. This provides a useful decomposition of returns on any non-cash investment into an excess return plus a cash return. In other words, the position itself can be viewed as an underlying position seeking excess return plus an equivalent cash position in that currency seeking a cash return. The net impact of a currency hedge is to substitute the cash return in the foreign currency by that in the domestic currency. It is important to note that the excess-return part, defined as market return above the risk-free rate, remains the same. In Table 6.4 we rework the asset-allocation examples using the excess-return approach, and the results are listed in Table 6.7. Comparing Tables 6.4 and 6.7, we see that both approaches give the same asset allocations. The attribution process is very similar to that of the currencysurprise approach detailed in Table 6.6. First, however, we need to compute the relevant excess returns and cash returns, and these are given in Table 6.8. The attribution calculation is given in Table 6.9. Table 6.8 Excess returns and cash returns calculated for excess-return approach from data in Table 6.5(b)
Nikkei
Unhedged (local currency)
Risk-free rate (USD)
Excess return
Cash return (USD)
(a)
(b)
(g)=(a)-(b)
(h)=(b)+(d)
14.00%
9.00%
5.00%
4.00%
S&P
8.00%
2.00%
6.00%
2.00%
USD Cash
2.00%
2.00%
0.00%
2.00%
5.25%
2.90%
Benchmark returns Underlying
314
CO2_with_index_2nd_amends.indd 314
23/05/2011 14:09
Benchmark Design and Performance Measurement
Table 6.9 Return attribution using the excess-return approach Underlying Nikkei
S&P
USD cash
Currency
Total
−0.06%
−0.14%
−0.20%
=(70%−45%)*(5%−5.25%)
=(10%−22.5%)*(4%−2.90%)
−0.23%
−0.11%
=(20%−50%)*(6%−5.25%)
=(90%−77.50%)*(2%−2.90%)
−0.26%
0.00%
−0.26%
−0.25%
−0.80%
−0.34%
=(10%−5%)*(0%−5.25%)
Total
−0.55%
The results in Tables 6.6 and 6.9 are identical. We should mention here that the excess-return approach can easily be applied to situations where more than two currencies are involved – one only has to deal with an array of excess returns plus a matrix of cash returns in different currencies. For the currency-surprise approach we need a matrix of underlying returns hedged into different currencies plus a matrix of pairwise currency surprises – a little more difficult to keep track of. Many studies of the effectiveness of currency hedging make unfortunate mistakes on precisely this point. 6.7 HOLDING-BASED BENCHMARKS In this section we give stylised examples to show how some commonly seen benchmarks might be constructed. These benchmarks are holding-based as opposed to performance-based; the latter will be discussed later in this chapter. There are many holdingbased benchmarks (which are in essence historical performances) in overlay managers’ pitch books and many articles on the benefit of currency overlay. Some of the results are simply too good to be true. And they probably are. Most offer no detailed explanation on how the figures are obtained. As we have demonstrated in Chapter 2, Section 2.2, and in Section 6.4, the calculation method may 315
CO2_with_index_2nd_amends.indd 315
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
have a critical bearing on the final results. To compute a benchmark is hardly rocket science, and the following examples are all constructed using easily obtainable data. In all our examples we choose three currencies – US dollar, sterling and Japanese yen – but the result can easily be modified to include more currencies. The base currency is the US dollar and the initial weightings are equally 33.33% (these are kept fixed for convenience, with no adjustment for asset drift). We assume that the equity indexes have been adjusted for dividends. Tables 6.10 and 6.11 contain the raw data and calculated raw returns used in subsequent sections. 6.7.1 Passive benchmarks The following benchmarks are passive because they all have fixed hedge ratios. 6.7.1.1 Unhedged Most publicly available equity indexes are unhedged. No adjustment is made for currency hedging in an unhedged index. Returns denominated in the base currency have two components: local currency returns and spot currency returns, the latter consisting of currency surprise (hedgeable) and forward premium (unhedgeable). Our way of computing unhedged returns may seem cumbersome, as they can be obtained directly from the returns on three indexes, all denominated in the US dollar, namely, R1, R8 and R9 in Table 6.11. However, calculating returns in this way has two distinct advantages: it lends itself directly to other passive hedge ratios, as shown in subsequent sections; and the equity return column, P2 in Table 6.12, which shows the underlying returns with forward points included, should be used as the index for the underlying manager, as argued in Section 6.3.
316
CO2_with_index_2nd_amends.indd 316
23/05/2011 14:09
CO2_with_index_2nd_amends.indd 317
Table 6.10 Raw data used in holding-based benchmark calculation examples S&P($)
FTSE(£)
Nikkei(¥)
USD/ GBP
USD/JPY
$ LIBOR
£ LIBOR
¥ LIBOR
FTSE($)
Nikkei($)
L1
L2
L3
FX1
FX2
I1
I2
I3
L2/FX1
L3/FX2
754
6,269
19,540
0.6191
107.37
5.89%
6.11%
0.09%
10,126
181.99
29/02/00
739
6,233
19,960
0.6336
110.14
5.92%
6.16%
0.13%
9,837
181.21
31/03/00
815
6,540
20,337
0.6276
102.66
6.13%
6.05%
0.10%
10,421
198.11
28/04/00
781
6,327
17,974
0.6435
108.15
6.29%
6.19%
0.09%
9,832
166.18
31/05/00
762
6,359
16,332
0.6664
107.72
6.65%
6.12%
0.08%
9,542
151.61
30/06/00
790
6,313
17,411
0.6588
106.02
6.64%
6.07%
0.15%
9,583
164.22
317
Benchmark Design and Performance Measurement
31/01/00
23/05/2011 14:09
S&P($)
FTSE(£)
Nikkei(¥)
USD/ GBP
USD/JPY
£−$ Libor ¥−$ Libor FTSE($)
Nikkei($)
R1
R2
R3
R4
R5
R6
R7
R8
R9
29/02/00
−2.09%
−0.57%
2.13%
−2.32%
−2.56%
−0.02%
0.48%
−2.89%
−0.43%
31/03/00
9.85%
4.82%
1.88%
0.95%
7.04%
−0.02%
0.48%
5.76%
8.91%
28/04/00
−4.21%
−3.31%
−12.35%
−2.50%
−5.22%
0.01%
0.50%
−5.81%
−17.57%
31/05/00
−2.58%
0.50%
−9.58%
−3.50%
0.40%
0.01%
0.52%
−2.99%
−9.18%
30/06/00
3.70%
−0.74%
6.40%
1.16%
1.59%
0.04%
0.55%
0.42%
7.99%
31/01/00
£−$ Libor and ¥−$ Libor are differences between Libor interest rates for currencies indicated
CURRENCY OVERLAY – SECOND EDITION
318
CO2_with_index_2nd_amends.indd 318
Table 6.11 Calculated raw returns for holding-based benchmark examples
23/05/2011 14:09
Benchmark Design and Performance Measurement
Table 6.12 Passive holding-based benchmarks: unhedged benchmark Unhedged return
Equity return
Currency surprise
Cumulative unhedged return
P1
P2
P3
P4
29/02/00
−1.80%
−0.03%
−1.78%
31/03/00
8.17%
5.67%
2.51%
8.17%
28/04/00
−9.20%
−6.46%
−2.74%
−1.03%
31/05/00
−4.92%
−3.71%
−1.21%
−5.94%
30/06/00
4.04%
3.32%
0.72%
−1.90%
31/01/00
P1=P2+P3; P2=[R1+(R2+R6)+(R3+R7)]/3 and P3=[0+(R4-R6)+(R5-R7)]/3. P4 is a simple cumulative arithmetic sum of returns to date. There are many other ways of calculating cumulative returns, throughout the examples given we choose to use the simple methods.
6.7.1.2 Fully hedged (unitary) The return for each country in the fully hedged benchmark consists of the sum of the local asset market return and the currency premium. The currency-surprise component of the spot currency return is eliminated, as shown in Table 6.13. Table 6.13 Passive holding-based benchmarks: fully hedged (unitary) benchmark Hedged return
Equity return
Currency return
Cumulative hedged return
P5
P6
P7
P8
31/01/00 29/02/00
−0.03%
−0.03%
0.00%
31/03/00
5.67%
5.67%
0.00%
5.67%
28/04/00
−6.46%
−6.46%
0.00%
−0.79%
31/05/00
−3.71%
−3.71%
0.00%
−4.50%
30/06/00
3.32%
3.32%
0.00%
−1.18%
P5=P6=P2 and P7=0.
319
CO2_with_index_2nd_amends.indd 319
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
6.7.1.3 Partially hedged A partially hedged benchmark, shown in Table 6.14, is a weighted average of unhedged and fully hedged benchmarks. If we use a 50% hedge ratio, the returns are the average of the unhedged and fully hedged benchmarks. Table 6.14 Passive holding-based benchmarks: partially (50%) hedged benchmark 50% hedged return
Equity return
Currency return
Cumulative 50% hedged return
P9
P10
P11
P12
29/02/00
−0.92%
−0.03%
−0.89%
31/03/00
6.92%
5.67%
1.25%
6.92%
28/04/00
−7.83%
−6.46%
−1.37%
−0.91%
31/05/00
−4.31%
−3.71%
−0.60%
−5.22%
30/06/00
3.68%
3.32%
0.36%
−1.54%
31/01/00
P9=P10+P11; P10=(P2+P6)/2; P11=(P3+P7)/2.
6.7.2 Dynamic benchmarks When the hedge ratio is not fixed, the benchmark is referred to as dynamic. Here we consider a few such benchmarks, albeit the simpler ones, used by fund managers. In practice, many dynamic benchmarks incorporate research results similar to those discussed in Chapters 4 and 7. One type of dynamic benchmark not included here is the CAPM-optimised benchmark discussed in Chapter 5, Section 5.4. As argued there, we consider that this type of benchmark is too volatile in the ex ante setting and therefore unsuitable for practical use. 6.7.2.1 Momentum benchmark A momentum benchmark incorporates some kind of trend-following rules. Here we assume a simple filer rule: if last month’s actual currency return is positive, then do not hedge; otherwise, hedge fully. The benchmark can be constructed as shown in Table 6.15. 320
CO2_with_index_2nd_amends.indd 320
23/05/2011 14:09
Benchmark Design and Performance Measurement
Table 6.15 Dynamic holding-based benchmarks: momentum benchmark Hedging decision
Portfolio return
Equity return
Currency return
Cumulative portfolio return
GBP
JPY
P13
P14
P15
29/02/00
Hedge
Hedge
31/03/00
Unhedge
Unhedge
5.67%
5.67%
0.00%
5.67%
28/04/00
Hedge
Hedge
−9.20%
−6.46%
−2.74%
−3.53%
31/05/00
Hedge
Hedge
−3.71%
−3.71%
0.00%
−7.24%
3.32%
3.32%
0.00%
−3.92%
P16
31/01/00
30/06/00
P13=P14+P15; P14=P2; If a Hedging Decision=hedge, then currency return is zero, otherwise it takes on unhedged currency returns.
Table 6.16 Dynamic holding-based benchmarks: carry benchmark Hedging decision
Portfolio return
Equity return
Currency return
Cumulative portfolio return
GBP
JPY
P17
P18
P19
29/02/00
Unhedge
Hedge
31/03/00
Unhedge
Hedge
4.90%
5.67%
0.32%
5.99%
28/04/00
Hedge
Hedge
−6.13%
−6.46%
−0.84%
−1.30%
31/05/00
Hedge
Hedge
−3.71%
−3.71%
0.00%
−5.01%
3.32%
3.32%
0.00%
−1.69%
P20
31/01/00
30/06/00
6.7.2.2 Carry benchmark Another popular benchmark explicitly incorporates forward bias. This benchmark is sometimes known as value benchmark. However, some people also term a rule based on the fair value of the currency (for example, implied by PPP) against the market exchange 321
CO2_with_index_2nd_amends.indd 321
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
rate as a value benchmark. In this book, we use “value benchmark” to mean the latter. A very simple version for a carry benchmark might be: hedge currency exposure if the foreign currency forward trades at a premium and leave it open if it trades at a discount. In our example the yen always trades at a premium, so yen exposure will be hedged, whereas sterling initially trades at a discount and then at a premium. The hedging decisions and the benchmark are shown in Table 6.16. 6.7.2.3 Fundamental benchmark For a fundamental benchmark the hedging decisions are made on the basis of signals from fundamental data. There are endless possible choices of which fundamental data to use and, indeed, how it is used. One of the common choices use fair-value indicators such as PPP or REER (real effective exchange rate) against the actual exchange-rate levels. Here we construct a very simple example using the OECD Composite Leading Indicators (CLIs).9 We look at the monthly changes in CLIs for the US, the UK and Japan. The hedging decision is taken as follows: if the change of UK CLIs for a particular month is higher than that for the US (sterling is likely to appreciate), the exposure is left unhedged; and vice versa. The benchmark is detailed in Table 6.17. 6.7.2.4 Dynamic hedging benchmark A dynamic hedging style aims to replicate the payoff of an option by hedging a portion of the exposure that is equivalent to the option delta. Delta can be loosely interpreted as the likelihood of option protection being needed. We will discuss option replication in more detail in Chapter 7, but here is an example of how a dynamic hedging benchmark can be constructed. In the example in Table 6.18 we choose two six-month put options on sterling and the Japanese yen against the US dollar, assuming implied volatilities of 9% and 12% for dollar/sterling and dollar/yen, respectively. Note that, if you want to compare the performance of the dynamic hedging benchmark with that of a purchased-option benchmark, the option premium should be explicitly incorporated.
322
CO2_with_index_2nd_amends.indd 322
23/05/2011 14:09
CO2_with_index_2nd_amends.indd 323
Table 6.17 Dynamic holding-based benchmarks: fundamental benchmark Composite Leading Indicators (CLI)
Hedging decision
Portfolio return
Equity return
Currency return
Cumulative portfolio return
USD
GBP
JPY
GBP
JPY
P21
P22
P23
P24
31/01/00 −0.19
−0.29
−0.05
Hedge
Unhedge
31/03/00
−0.35
−0.32
−0.49
Unhedge
Hedge
28/04/00
−0.22
−0.51
0.01
Hedge
31/05/00
0.75
−0.46
0.05
Hedge
30/06/00
−0.53
0.08
−0.37
Source: CLI data are from OECD
7.85%
5.67%
2.19%
7.85%
Unhedge
−7.29%
−6.46%
−0.84%
0.56%
Hedge
−3.75%
−3.71%
−0.04%
−3.19%
3.32%
3.32%
0.00%
0.13%
323
Benchmark Design and Performance Measurement
29/02/00
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
Table 6.18 Dynamic holding-based benchmarks: dynamic hedging benchmark Hedge ratio
Portfolio return
Equity return
Currency return
Cumulative portfolio return
GBP
JPY
P25
P26
P27
P29
31/01/00 29/02/00
47%
47%
31/03/00
42%
19%
7.00%
5.67%
1.33%
7.00%
28/04/00
59%
43%
−8.49%
−6.46%
−2.03%
−1.49%
31/05/00
81%
43%
−4.21%
−3.71%
−0.50%
−5.70%
3.59%
3.32%
0.27%
−2.11%
30/06/00
6.8 ADJUSTMENT STYLE We have already discussed different approaches in hedging adjustment: whether to hedge against actual exposures, or benchmark exposures, or hedge to benchmark neutral. Here are three numerical examples that illustrate how these choices can be incorporated into benchmarks. In all examples, we assume that the benchmark hedge ratio is 50%. 6.8.1 Hedging actual exposure The currency-overlay manager’s (moving) target is the actual exposure of the fund, which consists of the benchmark exposure and the active exposure. In the example shown in Table 6.19, the benchmark underlying position is listed in Column 1 and the currency-overlay manager’s benchmark position is in Column 5. The manager will make sure that the final position in Column 9 represents the hedging positions needed for the active position in Column 2. The adjustment needed in Column 8 is driven by the active positions in Column 2. 6.8.2 Hedging benchmark weighting The currency-overlay manager’s target is simply the benchmark exposure of the fund. As shown in Table 6.20, currency positions are derived from the benchmark underlying exposures and the hedge ratio. The extra exposure portion in Column 7 created by active positions is ignored. No adjustment will be made with respect to the active positions taken by the underlying manager: this is shown in Column 8. 324
CO2_with_index_2nd_amends.indd 324
23/05/2011 14:09
CO2_with_index_2nd_amends.indd 325
Table 6.19 Overlay adjustment: hedge actual exposure Underlying security
Currency selection
Average market weight
Average currency weight
Active
Difference
Benchmark underlying exposure
Hedge @ 50%
Hedged benchmark
Actual
Adjustment
Final
(1)
(2)
(3)=(2)−(1)
(4)=(1)
(5)=−50%x(1)
(6)=(1)+(5)
(7)=(6)+(3)
(8)=−50%x(3)
(9)=(7)+(8)
Euro
20.0%
25.0%
5.0%
20.0%
−10.0%
10.0%
15.0%
−2.5%
12.5%
Japan
20.0%
20.0%
0.0%
20.0%
−10.0%
10.0%
10.0%
0.0%
10.0%
UK
10.0%
10.0%
0.0%
10.0%
−5.0%
5.0%
5.0%
0.0%
5.0%
US
50.0%
45.0%
−5.0%
50.0%
25.0%
75.0%
70.0%
2.5%
72.5%
100.0%
100.0%
100.0%
325
Benchmark Design and Performance Measurement
Benchmark
23/05/2011 14:09
Underlying security
Currency selection
Average market weight
Average currency weight
Benchmark
Active
Difference
Benchmark underlying exposure
Hedge @ 50%
Hedged benchmark
Actual
Adjustment
Final
(1)
(2)
(3)=(2)−(1)
(4)=(1)
(5)=−50%x(1)
(6)=(1)+(5)
(7)=(6)+(3)
(8)=0.0%
(9)=(7)+(8)
Euro
20.0%
25.0%
5.0%
20.0%
−10.0%
10.0%
15.0%
0.0%
15.0%
Japan
20.0%
20.0%
0.0%
20.0%
−10.0%
10.0%
10.0%
0.0%
10.0%
UK
10.0%
10.0%
0.0%
10.0%
−5.0%
5.0%
5.0%
0.0%
5.0%
US
50.0%
45.0%
−5.0%
50.0%
25.0%
75.0%
70.0%
0.0%
70.0%
100.0%
100.0%
100.0%
CURRENCY OVERLAY – SECOND EDITION
326
CO2_with_index_2nd_amends.indd 326
Table 6.20 Overlay adjustment: hedge benchmark exposure
23/05/2011 14:09
CO2_with_index_2nd_amends.indd 327
Table 6.21 Overlay adjustment: hedge back to benchmark exposure Underlying security
Currency selection
Average market weight
Average currency weight
Active
Difference
Benchmark underlying exposure
Hedge @ 50%
Hedged benchmark
Actual
Adjustment
Final
(1)
(2)
(3)=(2)−(1)
(4)=(1)
(5)=−50%x(1)
(6)=(1)+(5)
(7)=(6)+(3)
(8)= − (3)
(9)=(7)+(8)
Euro
20.0%
25.0%
5.0%
20.0%
−10.0%
10.0%
15.0%
−5.0%
10.0%
Japan
20.0%
20.0%
0.0%
20.0%
−10.0%
10.0%
10.0%
0.0%
10.0%
UK
10.0%
10.0%
0.0%
10.0%
−5.0%
5.0%
5.0%
0.0%
5.0%
US
50.0%
45.0%
−5.0%
50.0%
25.0%
75.0%
70.0%
5.0%
75.0%
100.0%
100.0%
100.0%
327
Benchmark Design and Performance Measurement
Benchmark
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
6.8.3 Hedging to benchmark neutral In this arrangement, illustrated in Table 6.21, the currency-overlay manager first neutralises any deviations caused by the active underlying positions relative to the underlying benchmark, as shown in column 3. They will then take additional currency-hedging positions according to the benchmark weighting and the passive hedge ratio. 6.9 OTHER PRACTICAL CONSIDERATIONS Many of the choices we discussed in Chapter 3 can and should be incorporated into the benchmarks, making the setup much richer and complex than the ones we discussed so-far. Below is a list of some of them in very generic terms. o Size of notional amounts: In all the examples in this chapter, we deal only with percentage returns computed from spot and forward exchange rates; a real benchmark will need to incorporate the size of exposures and hedging positions. o Constraints in the mandate: These need to be incorporated into the benchmark. o Rebalancing frequency and triggers: The notional amounts of exposures and hedging positions will be changed during rebalancing. o Timing of transactions and fixing arrangements: These would affect the exact foreign-exchange rates used for benchmark calculation. o Choice of hedging instruments. o Instrument tenor. o Transaction costs: Given the discussion on this topic in Section 2.9.2.2, it makes sense to incorporate them if possible in the benchmark. Regular reviews are necessary. o Proxies: These need to be explicitly included in the benchmark. o Cashflow: Incorporating the effect of cashflows is difficult to do because it is linked to cash allocation and underlying investment. Regular reviews may be sufficient.
1. The well-respected Association for Investment Management and Research (AIMR) makes the following statement on the subject: “While AIMR here does not make a recommendation in favour of the unhedged or the hedged benchmark, we do regard the determination of currency exposure in the benchmark as an important fiduciary responsibility” (Stannard et al, 1998.)
328
CO2_with_index_2nd_amends.indd 328
23/05/2011 14:09
Benchmark Design and Performance Measurement
2. See, for example, Grinblatt and Titman (1994) and Lehman and Modest (1987) for equity indexes, and Blake, Elton and Gruber (1993) for bond indexes. AIMR’s 1997 publication, “Performance Evaluation, Benchmarks, and Attribution Analysis”, contains articles written by both practitioners and academics on the subject, but only a few articles deal with currency questions specifically. 3. Aside from the traditional government bond indexes, Lehman Brothers, Salomon Smith Barney and Merrill Lynch have each developed new broad market indexes to accommodate the much-enlarged investment universe. Many of these new indexes are accompanied by their currency-hedged counterparts. Salomon Brothers’ (now part of Citigroup) currency-hedged indexes use one-month forward contracts, rolled every month, with a hedge ratio of 100%. 4. MSCI has developed currency-hedged indexes for its popular equity indexes, including the EAFE series, as well as for several emerging markets on a custom basis. 5. For a fuller definition see Stannard et al (1998); available on AIMR’s website at www.aimr. com/standards/pps/benchmark.html. Table 6.1 is adapted from the same source. 6. Note that in this example we assume for convenience that the actual returns are the same as the passive returns. In practice, they will not be the same, as the former include the returns from security selection (eg, stock picking) and active currency contribution (eg, timing). However, it is easy to incorporate these in the framework. For more details see Karnosky and Singer (1994). 7. Interested readers can redo the example using local returns for the underlying and spot-onspot returns for currencies and compare the result. 8. It is difficult to generalise this for derivative contracts, as they have many different features. 9. More information on Composite Leading Indicators can be found on the OECD’s website at www.oecd.org.
329
CO2_with_index_2nd_amends.indd 329
23/05/2011 14:09
CO2_with_index_2nd_amends.indd 330
23/05/2011 14:09
7
Overlay and Alpha–Beta Separation
7.1 INTRODUCTION The industry trend at the time of writing has been alpha–beta separation: rather than using an active-overlay programme with twin objectives of reducing risk and enhancing yield, an investor may be better off using a passive overlay to reduce risk and a separate pure-alpha programme to enhance yield. What do alpha and beta mean for currencies? Are active-overlay programmes inherently less risky than pure alpha programmes because the former is a form of hedging while the latter is a form of speculation? Does it make sense to use this alternative approach? In this chapter we take a closer look at these questions. Parts of this chapter were based on Xin and Rondouin (2009) and d’Hautefort and Xin (2006). 7.2 ALPHA AND BETA FOR CURRENCIES 7.2.1 Definitions of currency alpha and beta The concepts of alpha and beta used in the asset-management industry come from a CAPM-style regression equation typically done for performance-attribution purpose. If you regress the returns from a particular portfolio on a market index using a simple linear equation, you get three terms: the constant term (alpha), the sensitivity term (beta) multiplied by return on the market index, and a disturbance term. Beta denotes the sensitivity of the portfolio return to the market returns; beta times market return therefore represents the part of the portfolio return that can be explained by 331
CO2_with_index_2nd_amends.indd 331
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
market movements, sometimes also referred to, loosely, as “market return”, or “index return”, or “passive return”. In contrast, the alpha term represents the part of the portfolio return that is unrelated to the market movements, sometimes also called “pure-alpha return”, “absolute return”, or “active return”. The pure-alpha return is usually attributed the manager’s skills. The remaining part of the portfolio return, usually random, would go to the third term, the disturbance term. Hence the return from a portfolio can be disaggregated into three parts: a market-related part (beta), a skill-related part (alpha) and the random part. It then follows that an investor could organise an investment portfolio in the same fashion: a beta part, relating to various types of the market returns; and an alpha part, relating to different types of trading strategies through which managers can demonstrate their skills (Waring and Siegel 2005). The alpha concept can be readily transcribed into currency management: all currency returns from active trading can be viewed as pure-alpha returns. Active currency trading is nothing but a longshort strategy. The beta part is less straightforward, as there is no market index for currencies. Some people would call the returns on a portfolio of currency exposures from a global benchmark index currency beta. For example, the return on the currency exposures in the MSCI EAFE index is sometimes called the currency beta return in the same index. If you regress the actual currency return against this beta return, the alpha coefficient from the regression would represent the return from active management, but the beta coefficient is actually the benchmark hedge ratio subtracted from 100% – something different from other asset classes. Things get muddier still when some other people use currency beta to refer the returns from some active strategies, for instance, currency carry beta, currency momentum beta, and so on. This type of currency beta index has been very popular in recent years. 7.2.2 Beta dressed up as alpha and alpha transformed to beta When some researchers studied hedge fund returns, they regressed them against a host of market return indexes and discovered that a fair portion of hedge fund returns could have been explained by beta returns. Some commentators said part of the hedge returns were “beta dressed up as alpha”. Later, researchers devised more 332
CO2_with_index_2nd_amends.indd 332
23/05/2011 14:09
Overlay and Alpha–Beta Separation
complicated beta indexes that embody certain trading strategies, for example, the equity long short strategy and convertible arbitrage strategy. Then they regressed hedge fund returns against these beta indexes and, of course, there were further outcries of “beta dressed up as alpha”. With the introduction of various currency beta indexes, similar studies and similar claims started to emerge: that currency active returns were also “beta dressed up as alpha”. The proliferation of all sorts of currency beta indexes does indeed set the bar higher for all currency managers; it also offers investors more choices for investing in active currency strategies. One final note on this point: the currency beta indexes we discuss here actually resemble active returns rather than passive ones. The example given earlier involves monthly adjustment to the portfolio, albeit with some rules. Hence, in a sense, those indexes transformed currency alpha into currency beta. It may be a little confusing for investors, but should be a good thing, as beta is usually cheap and easily accessible. In this Chapter, when we talk about alpha–beta separation, beta means market-related currency movements, which could be hedged using a passive currency overlay. Later we will see that the alpha part could be implemented by beta indexes. 7.2.3 Historical background Historically, currencies were always an afterthought for institutional investors. It was only after international asset allocation and global security selection that they realised fluctuations can be caused by a factor that they did not want, namely currency. In fact, it can cause plenty of fluctuations at around a quarter of the total portfolio risk in a global equity portfolio and up to three-quarters in a global fixed-income portfolio. Hence investors in the 1980s started using passive currency overlay to control currency risk in their global investment portfolios. Having a passive currency overlay in isolation does have an undesirable effect on the cashflow management since you need to write out cash cheques when the base currency is depreciating. It is only natural to ask whether this unwelcome cash drain could be mitigated by having an active currency overlay where the hedge ratio can be reduced in anticipation of the decline in the base currency. That sounds like an active currency overlay. Indeed, the first active-overlay mandates were created during the big US dollar cycle in the 1980s to mid-1990s, as shown in Figure 7.1. 333
CO2_with_index_2nd_amends.indd 333
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
Figure 7.1 US dollar evolution against a basket of currencies US Dollar Index
160
US dollar weakening: Hedged US investors suffer large cash outflows at hedge maturity
150 140
US dollar strengthening: Unhedged US investors lose money versus hedged investors
130 120 110 100 90 80 70 60 1970
1975
1980
1985
1990
1995
2000
2005
2010
Source: Reuters Ecowin Pro
An active-overlay programme has dual aims of reducing portfolio volatility and enhancing portfolio returns, but it is not without problems. A strategy good for enhancing returns may not be optimal or even feasible from a risk-reduction viewpoint. Most of the active-overlay mandates are constrained by various limits, discussed in Chapter 3, Section 3.2.7, thus limiting the abilities of the manager to express their views and demonstrate their skills. This phenomenon is particularly observed when the mandate refers to a polar benchmark (ie, unhedged or 100% hedged), as the potential added value will be limited by the broad trend in the base currency. One other type of handicap for an active-overlay manager is the common constraints of “no leverage” in any exposure currencies. For example, there may be good opportunities in trading Australian dollar crosses but the overlay manager is not able to add much value due to the small weights of Australian securities in the underlying portfolio. Also, trying to compare the added value of several managers against different benchmarks and across different periods of time is a very tricky task. But handicaps faced by currency-overlay managers are not necessarily good reasons for investors to change a certain practice, in this case, active currency overlay. The recent industry trend of switching out of active currency-overlay programmes was more the result of something else, explained in the next subsection. 334
CO2_with_index_2nd_amends.indd 334
23/05/2011 14:09
Overlay and Alpha–Beta Separation
7.2.4 Industry trend The industry trend over the last decade has been, instead of an active currency-overlay programme, a combination of passive currency overlay and a currency pure alpha. The combination approach would deal with two objectives of currency management: risk reduction and return enhancement, separately – a passive currency overlay for risk reduction and a pure-alpha programme for return enhancement. It also partially addresses the shortcomings suffered by active currency-overlay programmes. Another, more important, reason, in our opinion, is investors’ increasing recognition that alternative investments, such as hedge funds, are an asset class on their own. Active currency management, by definition a long/short strategy, is one of the most “traditional” hedge fund strategies. It has been widely used for more than a decade; its alpha generation is consistent and possesses a low correlation with other hedge fund strategies. This trend is summarised in Figure 7.2. Figure 7.2 The evolution of currency risk-management approach
Risk Reduction
Passive Overlay Aim to reduce the currency impact on returns from an international portfolio Reduce tracking error from active asset allocation and security selections Apply a passive hedge ratio, which is usually between 0% and 100% Can employ an option replication strategy
Return Enhancement
Active Overlay Aim to maximise portfolio returns as well as reduce currency impact on an international portfolio Use one or more forecast methods (trading strategies) to predict currency movements Take currency positions relative to existing underlying investments (ie, active hedging)
Separate “Alpha” from “Beta”
Absolute Return Programme Aim to maximise returns from any portfolio, with or without international investments The risk reduction objective is met by having a passive overlay programme Use one or more managers who use different forecast methods Risk levels (and target returns) set by investors Can be funded or unfunded
But, for some investors, a doubt lingers: an active-overlay programme would use only the existing currency exposures in a portfolio. In other words, it would not add any extra currency positions, 335
CO2_with_index_2nd_amends.indd 335
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
hence it does not really add risk. But a currency-pure-alpha programme? By definition, it amounts to currency speculation, as the positions would have no reference to any underlying exposures. Surely it is much more risky compared to an active currency-overlay programme. This is the question we will try to address in the next section. 7.3 THE RATIONALE BEHIND PASSIVE OVERLAY PLUS PURE ALPHA Whereas many studies confirm that a suitably chosen passive hedge ratio can indeed reduce the overall portfolio volatility, it is not clear whether the active currency overlay increases or decreases the risk. Many investors assume that an active currency overlay reduces risk or, at least, does not increase the portfolio risk, because it works with the currency exposures that already exist within the portfolio. As a corollary to that assumption, a currency-pure-alpha programme, where active currency positions are taken with no reference to the underlying portfolio, could potentially add risk, because new exposures would be added to the portfolio. This section aims to investigate these questions and analyse whether different management styles used to conduct currencyoverlay programmes have an impact on the portfolio volatilities. Here we constructed a stylised equity portfolio with the following allocation: 60% US equity, 20% euro equity, 10% each for Japanese equity and UK equity. We chose MSCI equity indexes for the underlying equity price moves and we used daily data from December 1999 to December 2006. 7.3.1 Starting with a passive overlay We first look at the impact of different benchmark hedge ratios on the total portfolio volatility – total volatility includes both equity volatility and foreign-exchange risk. This will serve as the baseline case for later analysis on active currency overlay and currencypure-alpha programmes. Passive currency overlay is the simplest way to manage the currency risk of a portfolio. We look at the impact of passive currency overlay on our equity portfolio with hedge ratios between 0% and 100%. One-month forward contracts are used in line with common practice. Figure 7.3 plots the total volatility of the passively hedged 336
CO2_with_index_2nd_amends.indd 336
23/05/2011 14:09
Overlay and Alpha–Beta Separation
equity portfolio against passive hedge ratios for different base currencies. It is easy to observe that, apart from JPY-based portfolios, passive currency overlay did reduce the total portfolio risk over the period. An AUD-based portfolio is an interesting case, as the total portfolio volatility initially decreased with the higher hedge ratio, and the maximum benefit in terms of risk reduction was achieved with a passive hedge ratio between 30% and 40% and a corresponding risk reduction of 0.30% to 0.40%. If the hedge ratio was higher than 70%, the total portfolio volatility started to increase, an indication of the benefits of currencies as a source of diversification. From a single chart, it is difficult to generalise. But, repeating the exercise for different portfolios and over different periods, you get very similar results. We can conclude with reasonable confidence that passive currency overlay generally reduces portfolio volatility, though probably not by as much as many people would have thought. And the risk reduction achieved, if any, is dependent on the base currency. Figure 7.3 Total portfolio volatility for different passive hedge ratios and for different base currencies 22.0%
Total portfolio volatility
EUR based
GBP based
USD based
JPY based
AUD based
20.0%
18.0%
16.0%
14.0% 0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Passive hedge ratio Source: Overlay Asset Management, Reuters
337
CO2_with_index_2nd_amends.indd 337
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
7.3.2 Comparing active with passive overlay Historically, many investors perceive active currency overlay as the less risky approach to management of currency risks, with the potential benefit of making money. Here we simulate active currency-overlay performance using two different styles, which are the most popular styles used by overlay managers (Middleton 2005): a trend-following style to mimic the option replication and a carry-trade style. For the trend-following style we use a naïve double moving-average crossover model with two sets of parameters to encompass both short-term and long-term trends; for the carry trade we use a simple rule of going long high-yielding currencies against low-yielding currencies filtered. We varied the benchmark hedge ratio from 0% to 100%. Note that it is also necessary for the investor to specify a passive hedge ratio for an active currency overlay; the overlay manager will then take his active positions around this benchmark hedge ratio subject to a “no leverage” constraint. In order to compare the results of active overlay with corresponding passive overlay, in the five panels in Figure 7.4 we show the volatility spread between them for different base currencies. It is easy to see from the charts that active-overlay programmes, either trend-based or carry-based, are not necessarily less risky compared with passive overlay. In fact, in most cases, active-overlay programmes result in higher total portfolio volatility compared with their passive-overlay counterparts; this is true for both carry-based and trend-based overlay programmes. This may be disconcerting for those investors using active currencyoverlay programmes with a dynamic hedging approach: the investors may assume that their currency risk is well taken care of by a programme designed to focus on the wax and wane of their base currencies. But the simulation results here provide no such comfort. A trend-based active-overlay programme is not superior, in terms of reducing total portfolio risk, to a carry-based activeoverlay programme. And nor is it better than a straightforward passive-overlay counterpart. For the sake of saving space, we do not show the total portfolio volatility for various active-overlay programmes and base currencies, but such information can be easily recovered by combining Figures 7.3 and 7.4.
338
CO2_with_index_2nd_amends.indd 338
23/05/2011 14:09
Overlay and Alpha–Beta Separation
Figure 7.4 Volatility spread between trend/carry-based active-overlay programmes and passive overlay programmes for different base currencies 2.00%
Active overlay (Yield based)
Volatility spread
1.50%
1.00%
0.50%
0.00%
-0.50%
Active overlay (Trend based)
USD based
Active overlay (Yield based)
1.50%
Volatility spread
2.00%
Active overlay (Trend based)
1.00%
0.50%
0.00%
0%
10%
20%
30%
40%
50%
60%
70%
80%
-0.50%
90% 100%
0%
10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Passive hedge ratio 2.00%
Passive hedge ratio 2.00%
Active overlay (Trend based)
Active overlay (Yield based)
Volatility spread
1.50%
1.00%
0.50%
0.50%
0.00%
0.00%
-0.50%
10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
EUR based
1.00%
-0.50%
0%
Active overlay (Trend based)
GBP based
Active overlay (Yield based)
1.50%
Volatility spread
JPY based
0%
10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Passive hedge ratio
Passive hedge ratio 2.00%
Active overlay (Trend based)
AUD based
Active overlay (Yield based)
Volatility spread
1.50%
1.00%
0.50%
0.00%
-0.50%
0%
10%
20%
30%
40%
50% 60%
70% 80%
90% 100%
Passive hedge ratio
Sources: OAM, UBS, Reuters
Of course, you could argue that total portfolio volatility is only one measure of portfolio risk and that, admittedly, some risks are even desirable. That is true. For example, we could use the maximum drawdown as a risk measure and compare the different impact of passive- versus active-overlay programmes. However, maximum drawdown from an active-overlay programme is directly affected by the profit and loss of the overlay programme, therefore the analysis would become a joint study of the risk reduction and the profitability of the active-overlay programmes in question. We choose in this section to use a “purer” risk measure in order to highlight the fact that active overlay can be in fact more risky than corresponding passive overlay with the same benchmark hedge ratio. 339
CO2_with_index_2nd_amends.indd 339
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
7.3.3 Adding pure alpha In the case of active currency overlay, the hedge ratio is usually constrained between 0% and 100%. In contrast, in a pure-alpha programme there is no such restriction. But is it more risky? We redo the analysis in the previous section but with one more option added: a passive-overlay programme coupled with a currencypure-alpha programme. The CISDM CTA Asset Weighted Currency Index is used as a proxy of the currency-pure-alpha returns.1 We restrict the volatility target of the pure-alpha programme to be 2% per annum2 – a target risk level commonly seen in practice. Similar to Figure 7.4, we plot the total volatility spread versus passive currency overlay, and the results are shown in Figure 7.5. Figure 7.5 Volatility spread between active overlay, passive plus pure alpha and passive-overlay programmes for different base currencies 2.00%
Active overlay (Trend based)
1.00%
0.50%
Active overlay (Yield based)
1.00%
0.50%
0%
-0.50%
10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
0%
Passive hedge ratio 2.00%
Active overlay (Trend based)
Active overlay (Trend based) Active overlay (Yield based)
1.50%
Passive overlay + 2% Pure alpha
Volatility spread
Volatility spread
10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Passive hedge ratio 2.00%
GBP based
Active overlay (Yield based) 1.50%
1.00%
0.50%
0.00%
-0.50%
JPY based
Passive overlay + 2% Pure alpha
0.00%
0.00%
-0.50%
Active overlay (Trend based)
1.50%
Passive overlay + 2% Pure alpha
Volatility spread
Volatility spread
2.00%
USD based
Active overlay (Yield based) 1.50%
EUR based
Passive overlay + 2% Pure alpha
1.00%
0.50%
0.00%
0%
10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
-0.50%
Passive hedge ratio
10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Passive hedge ratio
2.00%
Active overlay (Trend based) Active overlay (Yield based)
AUD based
Passive overlay + 2% Pure alpha
1.50%
Volatility spread
0%
1.00%
0.50%
0.00%
-0.50%
0%
10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Passive hedge ratio
Sources: OAM, UBS, CISDM, Reuters
340
CO2_with_index_2nd_amends.indd 340
23/05/2011 14:09
Overlay and Alpha–Beta Separation
The result in Figure 7.5 confirms that passive overlay plus pure alpha is not necessarily more risky than active overlay. Sometimes the total volatility is higher and sometimes it is lower, depending on base currency, benchmark hedge ratio and alpha risk budget. However, it is worth noting two interesting aspects in Figure 7.5. o The incremental volatility from a pure alpha programme is less sensitive to the choice of benchmark hedge ratio. This is expected, as the position-taking in a pure-alpha programme is smoother and more uniform, whereas that in an active currency-overlay programme can be lumpy for the polar hedge ratios. o The incremental volatility from a pure-alpha programme is fully controllable, whereas in an active currency-overlay programme it is dictated by the currency composition of the underlying portfolio. We have focused only on the risk side of the equation up to now and did not take a look at returns from various active currencyoverlay strategies. It is difficult to make sensible return comparisons between active currency-overlay strategies, as they are heavily influenced by restrictions such as the passive hedge ratio and the underlying portfolio, particularly when we look at only a few base currencies with a sample period of seven years. However, a comparison on risk characteristics usually suffers less from biases introduced by the sample time frame and/or the base currencies. 7.3.4 Summary of active overlay versus passive overlay plus pure alpha To conclude, an active currency-overlay programme is usually more risky than its passive counterpart if the risk is measured by total portfolio volatility, irrespective the style. A passive-overlay programme plus a currency-pure-alpha programme is not necessarily more risky than a corresponding active currency-overlay programme. More and more investors are switching from active currency overlay to passive currency overlay plus a currency-pure-alpha programme partly in recognition of this issue: since the benefit of active currency overlay comes mainly from the gains (alpha) it can generate, why straitjacket the currency managers against making those gains? If we sever the link between an active currency mandate and the exposures derived from global asset allocation, the manager can take positions in a broad universe of currencies regardless of the underlying port341
CO2_with_index_2nd_amends.indd 341
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
folio exposures. This solution aims at achieving consistent and attractive returns that can be “transported” into the passively hedged strategy of the portfolio. After all, there does not need to be a direct link between the currency components of the portfolio and the sources of alpha generated by the multiple inefficiencies of the foreign-exchange market. 7.3.5 Key benefits of using a passive-overlay programme plus a currency-pure-alpha programme The key benefits of a passive-overlay programme plus a pure-alpha approach are as follows. o It allows the investor to choose the “best of breed” managers for two distinctive activities: the manager(s) with the best systems to deal with operational risks for the passive currency-overlay programme and the manager(s) with the best alpha-generating strategies (with track record to show) for the pure-alpha programme. o It allows the investor to benefit from the currency manager’s skills without restrictions brought about by the underlying portfolios. o It allows the investor to easily use more than one pure-alpha manager, typically with different styles for diversification purposes. o It allows the investor to change pure-alpha managers more easily. However, this approach does have several difficulties apart from regulatory restrictions on currency speculation, such as choosing the right alpha manager(s) and choosing the right passive hedge ratio; and the cost of the pure alpha manager may be high, especially if the selected manager does not add alpha over the long term. 7.4 THE BEST IMPLEMENTATION METHODS AND OTHER CONSIDERATIONS There are several decisions to make when implementing a passiveplus-pure-alpha approach to currency management, the most important being the choice of the passive hedge ratio. In our view, the passive hedge ratio should be set with the main objective of reducing portfolio risk to an optimal level. The passive hedge ratio can be strategically adjusted based on some major misalignment in economic fundamentals or very long-term trends, not short-term realised gains or losses. Once the passive hedge ratio has been selected, the investor has to answer the following questions: 342
CO2_with_index_2nd_amends.indd 342
23/05/2011 14:09
Overlay and Alpha–Beta Separation
target risk level; funded or unfunded; alpha manager or beta indexes; the choice of alpha manager(s); whether the same manager should be used to do both passive and pure alpha; and o other operational issues. o o o o o
There are of course some risks associated with adopting a passiveplus-pure-alpha approach, such as short-term gains and losses, as currency management is a medium-term investment and there may be periods lasting a couple of years where the returns are negative. Another consideration is the lack of industry standard or rating agencies to compare performance from different managers. Overexposure to a single management style or trading strategy could cause problems, as may “style drift” and the associated risk of creating undisciplined changes from an investor to managers based on their recent performance. Despite the issues highlighted above, we believe that the combination of a passive-overlay and currency-pure-alpha programme offers a more efficient alternative to the traditional active currency approach. As such, we believe the current market trend will continue to strengthen in this direction.
1. See http://www.isenberg.umass.edu/CISDM/. 2. This is done by scaling the returns series of CISDM Index so the annual volatility is 2%. Admittedly, this is done using ex post data. In practice, currency managers would have needed to dynamically adjust the leverage of the programme (or ex ante volatility) so that the ex post volatility is more or less on target.
343
CO2_with_index_2nd_amends.indd 343
23/05/2011 14:09
CO2_with_index_2nd_amends.indd 344
23/05/2011 14:09
8
Options, Dynamic Hedging and Overlay
8.1 INTRODUCTION There are numerous books on various aspects of Black–Scholes–Merton option-pricing theory at almost any level of sophistication: continuous-time finance, martingale representation, stochastic calculus, third- or even fourth-generation exotic derivative products in different segments of the capital market, and beyond. It is not our intention to squeeze them all into this already crowded space. The reason we nevertheless feel obliged to include a chapter on option-pricing theory is threefold. First, we wish to illustrate the link between one of the overlay styles – dynamic hedging (also known as option replication) – and option pricing. Second, it allows us to discuss a practical question asked by many currency managers: should currency options be used in an overlay programme? To do this, we need to take another look at the concept of implied volatility, which has its roots in option prices. Third, we follow up with a brief review of option skews and smiles – the fact that the implied volatility of an option differs for different strikes. We look at why skew exists and how it may be exploited for investment. The chapter closes with some observations on the option “Greeks”. The message we would emphasise in this chapter is that optionpricing theory performs one fundamental yet often neglected role: the quantification of risks, enabling insurance contracts such as options to be priced against these risks. In this context option-pricing theory is – whether options are actually used or not – very relevant to many aspects of risk management objective associated with currency overlay. 345
CO2_with_index_2nd_amends.indd 345
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
8.2 OPTION-PRICING THEORY AND DYNAMIC HEDGING 8.2.1 Option price is determined through dynamic hedging We start with an intuitive description of what is behind the complicated formulae used to price options. In the simplest scenario, a customer buys an option from a bank, either to hedge their own risk or as an investment vehicle. For many clients, buying the option is the end of the story. But for a bank it is only the beginning: the bank, now with a short position in that option, needs to hedge its risk. No big deal – banks have been doing this over many years for many types of risk. There are, however, two special and very important features about this hedge. o In all likelihood the hedge will need to be adjusted now and then, depending on how market prices change. And, in theory, it needs to be adjusted continuously – this is one of the reasons why sometimes people talk about “continuous-time finance”. Not only that, the adjustment takes place in a prespecified manner according to option-pricing theory. o The adjustment process has nothing to do with what the underlying might be in the future. In other words, it does not matter what we think the price might be in the future. The option can, therefore, be hedged (and priced) as if everyone took the riskneutral view. If we ignore some technical caveats, the above two features form the foundation of option-pricing theory and, with it, continuoustime finance. The practical relevance of option-pricing theory is mainly in the following two areas.1 o The option price is not just a price for an instrument with a funny-shaped payoff profile. It is a price plus a DIY guidebook. If you follow the rules in the guidebook you can, with some preparation, replicate what the option is designed to achieve by trading the underlying continuously. o Using the guidebook (or, rather, the principles in the guidebook), you can derive a price for a given derivative instrument linked to an underlying asset with a given distribution. In other words, option-pricing theory allows the practitioner to price risks and, with the help of the guidebook, to manage these risks. 346
CO2_with_index_2nd_amends.indd 346
23/05/2011 14:09
Options, Dynamic Hedging and Overlay
The binomial representation of option-pricing theory can be found in many works on the subject. We will apply it in the case of a hypothetical call option to illustrate the points mentioned above. Our example makes the following assumptions: o the US dollar/Japanese yen spot price is 100.00; o dollar and yen interest rates are zero (to make the calculation simple); o there are two possible spot movements in the next period: up to 120.00 or down to 90.00, with probabilities of 60% and 40%, respectively; and o a call option struck at 100.00 will pay the holder ¥10 if spot ends up at 110.00 and nothing if spot ends up at 90.00. Figure 8.1 Binomial representation of option-pricing theory Spot Process
Option Payoff
120.00
10.00
60%
10.00
60%
100.00
?%
V=?
40%
(S,C)
40% 90.00
Time 0
Hedging Process
Time 1
?% 0.00
Time 0
Time 1
0.00
Time 0
Time 1
What should the market price of this call option be? We will denote the value of the option as V. To replicate the position, let us assume that we hold a portfolio of spot dollar/yen positions and a cash position, denoted as S and C, noting that S represents the US dollar notional amount of the spot position.2 The spot process, option payoff and hedging process are illustrated in Figure 8.1. We want the value of the portfolio (S,C) to be the same as the option payoff at time 1, ie, Up: S × 120.00 + C =10.0 Down: S × 90.00 + C = 0.0
347
CO2_with_index_2nd_amends.indd 347
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
Solve the equations (without calculating the exact results) and we get: 10.00 − 0.00 120.00 − 90.00 10.00 − 0.00 × 120.00. C = 10.00 − 120.00 − 90.00 S=
Therefore, if we hold such a portfolio of S and C, we will exactly replicate the payoff of the option. And, by the argument of no arbitrage, the value of this portfolio should be the value of the option at time 0, namely, V = S × 100 + C =
10.00 − 0.00 10.00 − 0.00 × 100.00 + 10.00 − × 120.00 . 120.00 − 90.00 120.00 − 90.00
Rearrange the last equation and we get the following: 100.00 − 90.00 100.00 − 90.00 V = 1 − × 10.00 + × 0.00. 120.00 − 100.00 120.00 − 100.00
This looks suspiciously like a formula for taking expectations: the terms in brackets are probabilities and the terms outside are option payoffs. More amazingly, the “real” probability of whether the spot will go up or down does not even appear in the expression! A whole new branch of finance, after the seminal work of Black, Scholes and Merton, is devoted to studying the whys and hows of this equation. Luckily, intuition is sufficient for our purpose and our excursion more or less stops here. Interested readers can turn to the many books available in this field – for example, Hull (2002) for a balanced exposition of both theory and practice, and Baxter and Rennie (1996) for some mathematics. 8.2.2 A few intuitive takeaway points from the option-pricing theory This simple example tells us the following. o Options (or, more generally, all derivatives contracts) are priced at the expected value of their future payoffs, but the expectation is taken using a different probability set. o Since the probability has only the price terms (it should also have interest rate terms but here we assume these are zero), it does not depend on people’s attitude to risk. It is as if everyone were risk348
CO2_with_index_2nd_amends.indd 348
23/05/2011 14:09
Options, Dynamic Hedging and Overlay
neutral. Hence, some people refer to the probability as the riskneutral probability. o The method we use to price the option has one by-product. It not only tells us the value of the option in question, but it also tells us how to construct a portfolio of spot and cash to replicate the option. This is the DIY guidebook that comes with the optionpricing theory. Of course, in the real world prices do not move up and down as in our simplistic setting, but the binomial representation can easily be extended to multiple periods and multiple branches and it can be used as a powerful tool for pricing derivative instruments. One final remark: where is the implied volatility term? In a one-period example we can link the volatility to the size of the steps. For a more volatile currency pair the up and down movement should be correspondingly bigger, and vice versa. 8.2.3 Transaction costs and option replication3 The binomial representation of a dynamic hedge can be adapted to continuous time. However, the presence of transaction costs now becomes an important issue because even low transaction costs can become prohibitively expensive in continuous time, and it becomes impossible to construct a perfect hedge that matches the option payoff exactly. Under such practical constraints, we need to follow strategies that optimise a chosen objective (for example, lowest cost, closest match, easiest implementation) under a particular set of constraints. A natural solution is to abandon continuous rehedging and to hedge only when the underlying price moves by more than a predefined interval. This is the so-called “optimal band” approach.4 Alternatively, we can follow an “optimal time interval” strategy, in which rehedging takes place only at certain time points.5 In practice, a dynamic hedger is likely to employ a combination of the two: when the underlying moves by a certain amount and/or after a certain period since the last rebalancing – a process not too dissimilar to the rebalancing method used for passive currencyoverlay programmes (Chapter 3, Section 3.3.6). The time trigger is necessary because the option hedge may change over time even if the underlying price stays constant. 349
CO2_with_index_2nd_amends.indd 349
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
The topic of rebalancing is directly relevant for currency managers who follow a dynamic hedging approach, as they will have to make these practical choices when running an actual hedging portfolio. Optimisation techniques, such as those used in the studies listed in Notes 4 and 5, can be used to determine the desirable rehedging band and time interval under a set of constraints appropriate for the given circumstances. 8.2.4 Why do people use a dynamic hedging approach to currency overlay? As we have mentioned earlier in this book, an important style of overlay management is dynamic hedging, also known as portfolio insurance or option replication. Similar techniques have been used by equity investors for several decades.6 The idea is straightforward: you hedge more when prices move against you, until you are fully hedged; and, conversely, you hedge less when prices move in your favour. In the context of currency overlay, you increase hedge ratio when the base currency goes up, and you decrease when it goes down. The payoff of such a strategy is increasing when the price moves in one direction (because your long position in base currency increases), and zero when the price moves in the other direction (because you reduce your position until you are flat). In its purest form a dynamic-hedging programme is completely systematic in that it does not depend on how you anticipate the spot price movements in the future. It is easy to see the similarity between a portfolio insurance programme and the payoff profile of a put option on the foreign currency and call option on the base currency. In fact, if we extended the binomial exercise shown in Figure 8.1 to more than one period, we would see the near-exact correspondence between a dynamic-hedging programme and what a market maker does to replicate an option. This brings up an important question: why should you use portfolio insurance rather than an option? One obvious answer is that it is free – compared with a purchased-option strategy, portfolio insurance does not require an upfront premium. Here we make a digression on what is known as a “free option”, which is used by some overlay managers and investment banks to de350
CO2_with_index_2nd_amends.indd 350
23/05/2011 14:09
Options, Dynamic Hedging and Overlay
scribe the dynamic hedging approach to currency overlay. The “free option” gives the user an unlimited upside while flooring the downside at zero (adjusted for forward premium). How is it possible to achieve this? From our construction of the optionhedging strategy in the last section, in order to replicate the payoff profile of the option but with no money upfront (ie, C=0), there is only one way: you have to be able to foretell the next spot move – ie, have perfect foresight. Then you can go long or short according to whether the spot goes up or down and accomplish the feat of replicating the option for free. That’s the string attached to the “free option”. People use a dynamic hedging approach instead of options for a combination of reasons. o Biased option prices: It is a widely held belief that options are more expensive than they should be. Empirical evidence on this point suggests, though not conclusively, that they are. Using the dynamic hedging approach can capture this pricing bias. o Cash cost and non-cash benefit: Options have to be purchased, and when they are not exercised they represent a sunk cost. Of course, this is balanced out by the “opportunity benefit”, whereby the underlying currency exposure will be in-the-money when an option is out-of-the-money. However, in the real world the opportunity benefit may not be treated the same as the upfront cost. o Flexibility: It is much easier to reverse a portfolio of spot/forward contracts than one of options. o Operational simplicity: Many old systems are not designed to deal with option-type contracts where the payoff at expiry is contingent. Options are also a lot more complicated than spot/forwards and specialist knowledge is required to deal with such contracts. o Regulatory prohibition: Some countries have stringent regulatory requirements on what type of instruments an institutional investor, especially a pension fund, can and cannot deal in, and the use of option-type contracts could be excluded. What might be the undesirable aspects of adopting a dynamic hedging approach rather than a purchased-option alternative?
351
CO2_with_index_2nd_amends.indd 351
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
o Slippage: When the spot market is volatile or fast-moving, a dynamic hedger may need to perform a substantial adjustment to their hedging portfolio. But, when the market is volatile, the cost of dealing is high and the hedger may not be able to achieve the desired hedging levels. This can result in underperformance of a dynamic hedging portfolio relative to an option, especially during the time when the hedge is needed most. o Transaction cost: Rather than cross the bid–offer spread in an option, a dynamic hedger will have to cross the bid–offer spread in spot (also swap or forward).7 In theory, the bid–offer spread in an option will reflect the expected bid–offer spread in spot replication. But in practice the former is often lower as a result of competition and position netting by option market makers. o Uncertainty: Rather than accepting a certain upfront premium, a dynamic hedger may face the uncertain cost of providing insurance. This is partly reflected in the slippage, but also in the overall cost of hedging.8 o Administration: A dynamic hedger may need to do many deals and monitor spot movements constantly. Does a portfolio insurance programme actually reduce the foreign-exchange risk of the portfolio more than, say, a passive currency-overlay programme? We have already seen from the previous chapter that it probably does not. But what about the relative merit in term of risk reduction of this approach versus purchasing an option? Occasionally, there are articles on the desirable payoff profile of a purchased option as a hedge, usually written by some investment banks, recommending its use in overlay programmes. Indeed, if you purchase a put option on the currency risk of the portfolio over a certain period, the maximum currency loss is limited to the premium paid for the option. Clearly, it alters the risk/ return profile of the portfolio of that period. If the original return distribution is symmetrical and bell-shaped normal, the altered distribution will have a cut-off point to the left and a hump in the left centre, indicating a greater likelihood of small losses. Over the period, an option or a portfolio insurance strategy will eliminate or reduce the likelihood of large losses at the cost of an increased likelihood of smaller ones. However, this is true only for that period. 352
CO2_with_index_2nd_amends.indd 352
23/05/2011 14:09
Options, Dynamic Hedging and Overlay
Unfortunately, hedging is usually not a one-off business for most institutional investors. In a multi-period setup, option premiums accumulate over time, and the effect of the changed distribution is increasingly diluted over longer horizons. If the horizon is long enough, a purchased-option strategy is similar to a dynamic-hedging programme in terms of risk reduction. 8.2.5 Dynamic hedging and trend-following9 But is the portfolio insurance programme – which aims to replicate an unpaid-for option – free? It is not immediately evident what the hedging cost actually is in our one-period binomial example. In a multi-period setting, the same principle applies and the dynamic hedger adjusts their spot position. This is a costly business because, if spot has gone higher during the preceding period, the hedger will need to buy more at a higher price; conversely, if the spot has gone down, they will need to offload some of their hedge at a lower price. So our hedger is constantly buying high and selling low – not a smart move for making money. In fact, the cumulative loss the hedger makes should equate to the option premium if the option is fairly priced. Some proponents of this approach then say it will create a “cheap option”. It may sound a little strange initially if we say that the option-replication approach and a trend-following approach are mechanically very similar to each other: the former involves buying high and selling low, whereas the latter is meant to buy in anticipation of uptrend and to sell, downtrend. If we look a little deeper, a trend-following approach needs to identify trends from recent price moves: usually recent upward moves would generate a buying signal and vice versa. The option-replication method works in the same way: recent price moves would result in changes of hedge ratio, upward moves prompting more buying and downward, more selling. So, rather than dwell on the numerous benefits of “free option” or “cheap option” supposedly provided by the dynamic hedging approach, an investor should think about whether a trend-following approach would make sense for their active-overlay programme. Compared with a purchased-option programme, a dynamic-hedging programme works well when the market trends strongly in either direction, or does not move very much; but not so well if the 353
CO2_with_index_2nd_amends.indd 353
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
market whips around the same level – you simply do a lot of buying high and selling low for no material benefit. But, when option replication approach works, a trend-following approach would too. If we accept the (limited) evidence that (1) options are more expensive than they ought to be and (2) price moves exhibit trends, then a dynamic-hedging programme has the potential to improve portfolio performance relative to a purchased-option benchmark. Also, the spot market is much more liquid than the options market and it is that much easier (and with less price distortion) to execute a dynamic-hedging programme than an options alternative for large portfolios. However, claims made by some commentators that an optionreplication approach is just for risk reduction and therefore its performance in terms of relative returns should not matter are misleading: it is trend-following, it should be benchmarked as such. 8.2.6 Using dynamic hedging to protect a “floor” Some overlay programmes using dynamic hedging approaches include explicit targets based on features of currency options. Such an overlay programme typically tries to protect a “floor” over a given period of time. A floor is usually defined as the maximum currency loss suffered by the investment portfolio over one year. For example, over a given one-year period, the maximum currency loss, net of P&L from hedging transactions, should not be over 3% of the total value of the portfolio. Loss can be defined in either spot-on-spot terms or spot-on-forward terms (see Chapter 1, Section 1.3.1, for issues with these two definitions). Three per cent is roughly the premium an investor would have to pay for an option to insure their basket of currency exposures against losses. If the investor buys this basket option, then their maximum loss over the one year period is fixed at 3%.10 The aim of the overlay programme is to outperform this floor. Some investors might find the idea of a “floor” particularly appealing because it gives them a sense of security in knowing beforehand the worst outcome due to currency moves. As with the floor to protect the downside, many such overlay programmes also have explicit aims for the upside, defined in “participation”, namely, if the currency moves over a period have been 354
CO2_with_index_2nd_amends.indd 354
23/05/2011 14:09
Options, Dynamic Hedging and Overlay
favourable, then the overlay programme is expected to participate in a portion of these moves. Usually, the portion is set between 50% and 85%. This feature also comes directly from options: if the investor bought the basket option for one year, then they would participate in 100% of all favourable currency moves. Admittedly, the participation is at the cost of the upfront premium payment for the option. The overlay programme does not pay the upfront premium, hence the target participation is usually set lower than 100%. Again, this feature is appealing, because it allows the investor to benefit from a portion of favourite currency moves. A currency overlay programme using a dynamic hedging approach should be able to achieve the two objectives over any given year as long as profits and losses are properly defined, currency volatility is not too high and correlation between currencies does not misbehave. However, as argued earlier in this chapter, hedging is an ongoing process, and a favourable pay-off profile (limited loss, unlimited gain) will be eroded over time, because, more often than not, such a programme would generate a loss over any given year by “selling low and buying high”, and the losses could accumulate quickly. On an ongoing basis, the profitability of such a programme is still driven by the profitability of a trend-following strategy. Moreover, neither the “floor” nor the “participation” can be guaranteed by currency-overlay managers. In a rapidly moving market, a floor could be broken. 8.2.7 The redundancy of options One question often asked by fund managers is whether it is necessary to use options, be they vanilla or the exotic varieties, in an overlay programme. From a theoretical viewpoint, under the classical Black–Scholes–Merton assumptions an option is a redundant asset. In Section 8.2 we demonstrated in a binomial setting that the payoff of an option can be replicated by trading spot. It can be proven that the same principle can be applied in continuous time: you can replicate the payoff profile of any option by continuously trading spot. If this is indeed the case, there would no reason for fund managers to use options and pay something extra for something they can construct themselves. This assertion leads to the following questions. 355
CO2_with_index_2nd_amends.indd 355
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
o Are options still redundant, theoretically and practically, when the classical assumptions are violated? o Is the option price justified? Namely, do end-users pay over the odds for what they get in the options markets? Many end-users dismiss the use of options by saying that they are too expensive for what they are. Cynics may further add that end-users are probably net buyers of options and that market makers – typically investment banks – raise prices accordingly to profit from this situation. o Is the bid–offer spread for options, which can be some magnitudes larger than that for a spot transaction, justified by the cost of constructing them? The second question relates to the efficiency of the currency options market, and we defer discussion of this until later in the chapter. The third question has not been systematically studied. But, given the competitiveness of the market and the fact that bid–offer spreads have gone down significantly in recent years, we think the answer to it is probably yes. Here we will take a look at the first question. Black, Scholes and Merton made a series of assumptions to derive the well-known option-pricing formula. When these assumptions are not met in practice – which typically they are not – the pricing formula needs various adjustments. Some of these are cosmetic: for example, interest rates are deterministic but not constant; volatilities are deterministic but not constant; and so on. Under these variations, the DIY rules in the guidebook are preserved, the payoff of the option can still be replicated risklessly and so options remain redundant. But other violations of the assumptions are more serious: incomplete underlying market, stochastic volatilities, transaction costs, etc. In these cases, risk-free replication can no longer be achieved and its place is taken by somewhat risky replications (usually through some sort of fancy optimisation). When the replication is no longer riskfree, the option will not be redundant: it cannot be manufactured by trading the underlying without incurring some sort of cost and/or inaccuracy. Its price will be different for people with different risk aversions and its market price will be set at a point where buyers’ interests equate to the sellers’. Since violation of the Black–Scholes assumption is the market norm, we can conclude that options are 356
CO2_with_index_2nd_amends.indd 356
23/05/2011 14:09
Options, Dynamic Hedging and Overlay
probably not redundant. However, this leaves one further question: to what extent are they not redundant? Are options only marginally different from what can be replicated by trading the underlying, and therefore can be ignored in practice, or are they completely different and should be viewed as an essential tool in the suite of investment tools? A direct approach to this question is difficult but there are two indirect empirical routes. One looks at whether the introduction of an option alters the price behaviour of the underlying and, if it does, then we can infer that options are not redundant. The other looks at the empirical returns and risks of options to see whether they are significantly different from those of the underlying. Investigation along the first route has been sporadic and has concentrated mainly on stock options.11 The findings usually support the notion that the introduction of options does significantly affect the underlying prices and therefore stock options are not redundant. The reason is often attributed to the incompleteness of the stock market. The currency market is significantly different from the stock market owing to its huge liquidity and standardised underlying assets. Connolly (1996) looked at the impact of options on currency futures and found no significant price effect on foreignexchange futures. He argued that the existence of other foreign-exchange derivatives contracts (over-the-counter options on spot, for example) are probably close substitutes and that the new options on futures are probably redundant. Chan and Lien (2002) examined the redundancy of options on currency futures by looking at the transmission of information between foreign-exchange futures and spot markets. They found that after the introduction of options the instantaneous feedback between spot and futures markets improves drastically. They also found evidence that options play a leading role in the transmission of information in currency markets. The second empirical approach looks at whether options provide a risk/return profile that is distinct from spot foreign exchange, or, indeed, any other asset class. To isolate the impact of spot moves on options, we need to look at volatility alone. Is volatility a separate risk class? And, if it is, is volatility risk priced in a CAPM sense? The academic methods used to answer the above questions are usually convoluted. In summary, the existing evidence from different markets seems to suggest that: 357
CO2_with_index_2nd_amends.indd 357
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
o volatility is not constant; o options, both at-the-money (ATM) and out-of-the-money (OTM) types, do exhibit a distinct risk/return profile and therefore should be considered as a separate risk class; and o volatility attracts a negative risk premium – ie, long volatility will, on expectation, produce a loss; in other words, options are more expensive than they should be (Jackwerth and Rubinstein 1996). Generally speaking, the existing research on both equity and fixedincome derivatives seems to indicate that the implied volatility is stochastic and cannot be spanned by underlying assets.12 Options, therefore, are not redundant. 8.2.8 Use of options for currency-overlay programmes On balance, the empirical evidence is supportive of the notion that currency options are not a redundant asset. However, it does not necessarily follow that a fund manager should use options in a portfolio. Yes, options can provide an additional source of return and/or be a powerful tool for risk management, but they can just as easily be a turbocharged value destroyer, as attested to by a long list of stories of derivatives going horribly wrong. In practice, for both active and passive-overlay programmes, the use of currency options is still limited to a few niche areas. Passiveoverlay programmes have very narrowly defined objectives that can be efficiently achieved by using forwards or futures. One of the goals of active-overlay programmes is to create an option-like payoff profile as shown in Figure 2.1. Purchasing options systematically is unlikely to attain that goal, as the accumulation of option premiums will quickly expand to the left tail of the distribution. Active-overlay programmes tend to have short investment horizons, which mean frequent adjustments to hedging positions, easier to do for forwards than for options. There are a few overlay managers that use options now and then on their active-overlay programmes, if allowed by the investors, and they usually adopt a fundamentaldiscretionary approach, with investment horizons between one and six months, longer than the systematic managers. Where options make a regular appearance is typically in some of the purealpha programmes offered by some currency-overlay managers. 358
CO2_with_index_2nd_amends.indd 358
23/05/2011 14:09
Options, Dynamic Hedging and Overlay
Trading options add two additional dimensions to spot/forwardbased strategies – timing and implied volatility – and therefore can expand the decision-making space. Our view is that whether to use options in a currency portfolio is dependent on a particular currency manager’s skill set and operational constraints. It is essential that the currency manager have the necessary expertise before adding currency options to a portfolio. Apart from some ad hoc transactions mainly for short-to-mediumterm insurance purposes, options are probably of limited appeal to investors using passive- or active-overlay programmes mainly for risk-control purposes. 8.3 SKEWS AND SMILES Before we consider the efficiency of the options market and potential trading strategies using currency options, we need to make a quick detour into the subject of options skews and smiles. The topic may sound too technical for investors not looking to use options on their currency-overlay programmes, but it is important nevertheless, as skews and smiles contain useful information about future distribution of foreign-exchange moves. Both terms refer to the phenomenon that implied volatility can differ for options with different strikes but otherwise identical features. There is no provision for this in the traditional Black–Scholes–Merton setup, as volatilities are assumed to be constant across time as well as strikes.13 In particular, “skew” refers to the observation that equal-delta puts and calls can be priced off different volatilities, whereas “smile” refers to the observation that volatility tends to increase when option strikes move away from the forward level. There is no simple formula or model that will allow somebody to construct a volatility smile from an at-the-money volatility quote because the smile is largely driven by demand/supply in the options market. Figure 8.2 shows how implied volatility, both ATM and OTM, and vanilla option premiums are linked. Using volatility, you can apply the standard Black–Scholes–Merton model to agree on the price for an option (links 1 and 4). Banks tend to have their own proprietary models to take into account market imperfections that are not dealt with in standard Black–Scholes–Merton (links 2 and 5). However, banks generally will not quote you option prices from 359
CO2_with_index_2nd_amends.indd 359
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
their own models. They instead adjust their implied volatility quote such that when the implied volatility is entered into a Black–Scholes calculator (links 1 and 4 again) it produces the same price as their proprietary models. The key point here is that there is no model that will give you OTM volatility from ATM volatility (links 3 and 6). Figure 8.2 Illustration showing how volatilities are linked to option premiums 1. Black-Scholes
ATM Volatility
ATM 2. Models, some take OTM vol into account
Premium
6. No model
3. No model, market-driven
4. Black-Scholes
OTM Volatility
5. Models
OTM Premium
In a Black–Scholes–Merton world, all options with the same maturity date would be priced on the same volatility. This makes intuitive sense, as there can be only one possible “actual” volatility for a spot process over a certain time period, and this value should not depend on the strike of the option. However, the market makers will change their prices for OTM options to reflect, among other factors, supply-and-demand conditions in the market. Generally speaking, demand outstrips supply for OTM options. A main reason for this is the high leverage for an OTM option. Rather like a lottery ticket, an OTM option costs very little, but, once the spot moves by a long way (admittedly this is a low-probability event), the potential gain as a percentage of the upfront premium (socalled leverage) is much larger than that of an ATM option. Therefore, there are usually more buyers than sellers for OTM options – a market makers’ nightmare. 360
CO2_with_index_2nd_amends.indd 360
23/05/2011 14:09
Options, Dynamic Hedging and Overlay
Other factors also contribute to the smile: o many corporate/institutional hedgers seek cheap insurance by purchasing OTM options; o expectation of spot movements will push up the OTM call relative to the OTM put, or vice versa (skewed smile); o empirical spot-return distributions display time-varying kurtosis and skew, which often leads to adjustments in OTM volatilities (the underlying distributions may not be normal or lognormal); and o last but certainly not least, with the growing volume of exotic options in the foreign-exchange market, market makers use OTM options to statically hedge exotic risks, which means that they are unwilling to short too many OTM options. This may in turn aggravate the smile. As one can imagine, no single model has emerged to capture all these factors, and it is very much down to demand/supply and market makers’ judgement to determine the magnitude of the smile at a particular point in time. 8.4 TRADING OPPORTUNITIES IN THE OPTIONS MARKET Chapter 4 was, in part, a survey of trading opportunities in the foreign-exchange spot market presented with the aim of helping investors to decide whether the opportunities are sufficiently attractive for them to trade currencies actively. The discussion would not be complete without a parallel review of the efficiency of the foreign-exchange options market. However, here the motivation is slightly different. As discussed in Sections 8.2.6 and 8.2.7, options can be a useful though not necessarily indispensable tool for an overlay manager, and the information on market efficiency is therefore directly useful only for managers/investors who choose to use options. But, in view of the importance of option-pricing theory in risk management in general, this section on trading opportunities in the options market may also be relevant for those fund managers who decide not to incorporate options in their portfolios. 8.4.1 Overview The question of efficiency in the foreign-exchange options market 361
CO2_with_index_2nd_amends.indd 361
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
can be loosely paraphrased as whether implied volatility, being the market consensus forecast of actual volatility, is an unbiased estimate of future actual volatility: alternatively, whether foreign-exchangeimplied volatility incorporates information such that excess return from taking any static or dynamic position in options does not exceed the transaction cost of trading based on that information. Two related questions are often asked by those who may not be familiar with foreign-exchange options markets: o Does implied volatility move a lot? o Does it move in a particular way that corresponds to spot move ments? The answer to the first question is unequivocally yes. For a currency manager, investing in spot can work in almost exactly the same way as investing in implied volatility.14 The second question – whether the moves in spot and implied volatility are correlated in some way – is much trickier to answer. On an intuitive level, implied volatility measures the two-way movement in spot in that we would expect implied volatility to increase when spot moves away from its mean. In an ideal world a scatter graph of changes in implied volatility against changes in spot should have a kidney shape. A correlation measure would be quite powerless in this situation, since, when spot went up, the correlation would be positive and, when it went down, negative. Models incorporating a changing volatility can be very convoluted and they do not often lead readily to an easily interpretable relationship between spot moves and implied volatility moves. Also, these models are not widely used in foreign exchange compared with, say, in equities or bonds. 8.4.2 Does implied volatility systematically overpredict actual volatility? Many fund managers claim that implied volatility systematically overpredicts actual volatility. If this is true, it makes options less attractive than spot or forwards as hedging instruments. But any suggestion that the latter are a superior hedging instrument is easily disputed, as one can just as easily sell implied volatility and make a statistical arbitrage profit if implied volatility is on average more expensive than actual.15 362
CO2_with_index_2nd_amends.indd 362
23/05/2011 14:09
Options, Dynamic Hedging and Overlay
A simple test of the claim of systematic overprediction would be to compare the implied and actual volatilities for a particular period. Here we choose six currency pairs in which OTC options are actively traded (US dollar/Japanese yen, euro/US dollar, euro/ Japanese yen, sterling/US dollar, euro/Swiss franc and euro/sterling), with three different time frames, one month, three months and twelve months. By definition, implied volatility is forwardlooking: the one-month implied volatility today is a forecast of spot volatility for the next month. On the other hand, one-month actual volatility today is computed from last month’s price data, ie, it is backward-looking. A fair comparison, therefore, would be to compare the one-month implied volatility from a month back with today’s one-month actual volatility. If we compute the historical difference between implied volatility and ensuring actual volatility for different base currencies, different maturities and different historical windows, we find that it is more likely that implied volatility will come out higher. While the spread of implied over actual volatility may be statistically significant, it is, however, not stable over time, as represented by standard deviation.16 Before jumping to the conclusion that the options market “overprices” the spot movement, we need to address a few questions: o Is implied volatility in the dataset accurate (or relevant)? o Is actual volatility calculated accurately (or could it be calculated accurately)? o Why should the spread be zero, anyway? A few factors may affect the results we have presented: o Bid–offer spread: We use mid-volatility in the test. Bid–offer spreads in these currency pairs are typically in the range 0.1–0.3%. o OTM volatilities: Out-of-the-money volatilities are not included in the test. It seems sensible to use a weighted average of all the volatilities rather than only ATMF volatility. However, including OTM volatilities would probably increase the spread and they are generally higher than ATM volatilities. o Return frequency: This is a very big topic on its own. To calculate actual volatility, should we use daily returns? Weekly returns? 363
CO2_with_index_2nd_amends.indd 363
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
Hourly returns? Or multiples or fractions of them? There is no definitive answer to the question – it usually depends on what you want this number for in the first place. One thing is clear, though: different return frequencies are likely to result in different estimates of the actual volatility.17 o Autocorrelation in returns: One of the reasons why the choice of return frequency is so important is the autocorrelation in returns. We can overestimate the actual volatility for a positively autocorrelated return series and underestimate it for negatively autocorrelated series, as illustrated in Figure 8.3. Our results are based on a simplistic setting and do not incorporate trading constraints. More robust tests are needed to validate the claim that options are overpriced. To launch a trading programme to exploit this, we would need to consider the risk profile of such a strategy as well. If we were to sell vanilla options routinely, the gains would be small and accumulate over time. But losses, when they appear, would be large. Figure 8.3 Auto-correlation and estimation of actual volatility
B
Positive autocorrelation. Overestimate volatility if calculating return from point A to B
A
Negative autocorrelation. Underestimate volatility if calculating return from point A to B
A
B
364
CO2_with_index_2nd_amends.indd 364
23/05/2011 14:09
Options, Dynamic Hedging and Overlay
8.4.3 Volatility and correlation forecasting Volatility forecasting has attracted an extensive array of research from both academics and practitioners.18 This partly reflects the important and diverse role of volatility in investment, security valuation, risk management and monetary policy making. For example: o Volatility directly impacts on the pricing mechanism of almost all options, which have become increasingly accepted as one of the main investment and risk-management vehicles; o Indirectly, volatility enters into the valuation of many types of assets and liabilities, through either CAPM or analysis of optionlike features in the assets and liabilities; o Volatility is a key ingredient in the portfolio-optimisation procedures widely used by investment managers; o Additionally, volatility may influence the mon etary decisionmaking process. The US Federal Reserve uses volatilities of stocks, bonds and currencies as one type of indicator, and the Bank of England also makes frequent reference to option-implied volatilities as a key financial variable; o Last but certainly not least, the 1996 Basel Accord links the VaR of banks and trading houses to their capital reserves and therefore makes volatility forecasting an integral part of risk-management systems for these financial institutions. There are two main strands of research in this area. One follows the seminal work of Engle (1982) and Bollerslev (1986) on Arch (autoregressive conditional heteroscedastic) modelling. In a nutshell, an Arch process posits that the volatility of the next period is closely linked to the volatility of the previous period adjusted by the latest (squared) movements of the underlying prices. It is therefore capable of reproducing the often-observed phenomenon of volatility clustering: large moves are likely to be followed by large moves and vice versa. A version of the Arch model called Garch (generalised autoregressive conditional heteroscedasticity) is widely used by foreign-exchange practitioners to forecast volatility. The other strand investigates the forecasting power of implied volatility, which is a traded commodity and regarded as the market-consensus forecast of future volatility. 365
CO2_with_index_2nd_amends.indd 365
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
The empirical evidence on how useful implied volatility is for forecasting purposes is at best mixed. Generally speaking, researchers find that it does contain some information about future actual volatility but that it is not necessarily a superior estimator to, say, historical volatility or an Arch forecast, and implied volatility is generally thought to be a biased estimator (see Section 8.4.2.19 In a widely quoted paper, Jorion (1995) examined implied volatility from options on foreign-exchange futures. His main finding was that implied volatility has substantially greater predictive power than time-series models (including Arch or Garch), even when the Arch models have the advantage of using ex post data to estimate its parameters. Jorion’s results are similar to those of West and Cho (1995). Weinburg (2001) confirmed Jorion’s result and also found that little discernible additional information is embedded in volatility smiles beyond the ATMF implied volatility. Neely (2002) used intra-day data and employed measures to correct a data estimation problem in Jorion’s study; he concluded that, on balance, implied volatility is a biased estimator of actual volatility. However, Neely’s study also contains two more interesting findings: implied volatility provides a better forecast of actual volatility computed using intra-day data; and time-series methods, including Garch, improve the forecasting power of implied volatility in an out-of-sample comparison. We should note that even with the combined help of implied volatility and statistical methods, only about a third of the variation in actual volatility is explained – in other words, these are not very good forecasts. Neely did not consider whether the apparent inefficiency can be exploited in a statistical arbitrage. The use of implied volatility to forecast correlation is a much less investigated field. Campa and Chang (1998a) found that implied correlation from foreign-exchange options outperforms timeseries-based correlation forecasts but that in some cases historical correlation also contains additional information. They used a criterion based on economic profit and loss, similar to that in Engle et al (1993), to compare the relative success of different forecasts. The method these authors employ is much closer to what a practitioner might use than those used by many academics. 8.4.4 Some volatility-based trading schemes Some financial instruments – for example, the forward volatility 366
CO2_with_index_2nd_amends.indd 366
23/05/2011 14:09
Options, Dynamic Hedging and Overlay
agreement (FVA) and the volatility–variance swap – allow investors to take positions directly in volatility without committing to a particular spot level or even with no exposure to directional spot movements. The volatility–variance swap is not so much a swap as a forward contract. It allows investors to profit from or hedge the risk of an increase or decrease in future currency volatility. Its payoff is computed by multiplying the notional face amount by the difference between the realised variance–volatility and the implied variance–volatility. Volatility–variance swaps can be used as more than just a speculative vehicle on the rise or fall in volatility. Currency-risk managers can use such contracts to hedge against unexpected future changes in volatility. According to a research note20 published in 2001 by FX Concepts, a New York-based overlay manager, some pension funds have started to look into the use of volatility products for both return-enhancement and risk-management purposes. The report notes that currency managers, predominately systematic players, have been slow to innovate in this area mainly due to a lack of expertise in options and a good volatility database. Both situations are improving quickly. Many banks now allow their clients to analyse and download volatility data, including skews and smiles, directly from their websites. The following subsections present three short examples in which investors adopt more traditional trading methods to trade volatility (or use volatility to trade spot) systematically. These represent actual trading schemes (the parameters used in some trading models can be different) as used by some market players. 8.4.4.1 Gamma scalping Gamma scalping offers a means to capture the spread between implied volatility and future actual volatility. If investors are of the view that the former is lower than the latter, they can buy the option and try to replicate a short-option position using currency spot/forward, just as we discussed in Section 8.2 (but bearing in mind that the example there is to replicate a long-option position). If the replication is successful, at expiry an investor will have the exact same position as if they were to short the option, which will neutralise the long-option position they entered upfront. They are left with two cash items: one 367
CO2_with_index_2nd_amends.indd 367
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
is the cash outflow from purchasing the option in the first place, and the other is the cumulative cash inflow, reflecting the profits of carrying the replication. This time the investor will always be selling high and buying low, so they will end up with a profit from their replication exercise – hence the term “scalping”.21 If the investor’s view is right, the net cashflow will be positive. 8.4.4.2 Selling wings volatility If implied volatilities are generally higher than the actual volatility, could you make a profit by selling OTM strangles? A strangle consists of one OTM put option and one OTM call option, usually with similar deltas. A number of market players are consistent sellers of such options. They are usually systematic in their approach, with steady notional amounts. Some years ago there were many more sellers of OTM volatilities – often on a discretionary basis and in large notional amounts – than there are now. The reason for using OTM options is that they can reduce the magnitude of the losses that a seller may suffer when selling volatilities. Also, due to the volatility smile, OTM volatilities are usually higher than the ATM counterpart. We are not aware of many publicly available studies on the effectiveness of this strategy, the main difficulty being to obtain a reliable dataset of OTM volatilities22. But the strategy is frequently used by many hedge funds and proprietary trading desks of banks. A UBS research note in 1999 provides some support for the existence of excess return, albeit not across different currency pairs (Xin 1999). 8.4.4.3 Trading spot using option skew as an indicator When the implied volatilities for equal-delta puts and calls that are otherwise identical differ, this is referred to as option skew. As we saw in Section 8.3, there are many reasons for option skew, and one of them reflects the market’s collective anticipation of a directional move on the spot market. If the spot is viewed as more likely to go up than down, call options are likely to be more expensive than the corresponding puts, as some market players may wish to purchase the call in anticipation of the spot move. We naturally ask the question: does option skew contain information about future spot movements that is not available else368
CO2_with_index_2nd_amends.indd 368
23/05/2011 14:09
Options, Dynamic Hedging and Overlay
where? Some empirical evidence is supportive of a yes answer. Campa, Chang and Reider (1998) investigated implied skewness in one- and three-month over-the-counter options on a number of different exchange rates between April 1996 and March 1997. They found that the direction of skewness was positively correlated with returns over the remaining length of the option. Malz (1997) also examined over-the-counter foreign-exchange rates using onemonth options from April 1992 through June 1996 and found that investors can earn excess returns (in a CAPM sense) by holding currencies whose option prices indicate positive skewness. Several banks have issued research notes that support the existence of excess return from this strategy. 8.5 SPEAKING GREEK – SOME CLOSING COMMENTS To summarise our views, we think that foreign-exchange options as financial instruments do offer a new avenue for currency management but that they may or may not be suitable for a particular fund depending on its circumstances. However, an understanding of foreign-exchange option-pricing theory is essential for sound risk-management practice. But how much understanding is enough? Do you have to know your delta, gamma, vega and rho and pepper your conversations accordingly? Or, if that is not enough, do you also have to know about d(vega)/d(volatility), “pin” risk, or the willow-tree implementation of the finite-difference method? Throughout the book we have tried to steer clear of such “geek” talk, which is common these days when you speak to foreign-exchange salespersons from banks and their financial engineers, risk advisers and structurers.23 The concepts behind some of this jargon are essential for a basic understanding of option-pricing theory, but the ability to drop technical names is, in our view, not. In many cases the fancy Greek terms have more mundane and commonsensical meanings and these are more important (and perhaps less confusing) to understand. So, next time a banker calls and tells you that gamma is bid or it is expensive to sit on a pile of rho, ask him what it means in terms of spot, forward, implied volatility and actual volatility.
369
CO2_with_index_2nd_amends.indd 369
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
1. The option-pricing theory, which has its origins in the PhD thesis “Théorie de la Spéculation” by the French mathematician Louis Bachelier in 1900, predated Einstein’s famous paper of 1905 on heat diffusion, which is based on a similar concept. However, it was 70 years before Black, Scholes and Merton came up with the ingenious argument on dynamic hedging and risk neutrality and the theory became one of the cornerstones of modern finance. 2. C should be a position in a bond or deposit, but since we assume that both interest rates are zero, we can ignore the time value of money here. 3. See Martellini (2001) for a survey on the subject. 4. See, for example, Hodges and Neuberger (1989), Constantinides (1997) and Whalley and Wilmott (1993). 5. Refer to Hoggard, Whalley and Wilmott (1994) for fixed time intervals and Henrotte (1993) or Grannan and Swindle (1996) for variable time intervals. 6. See, for example, Asay and Edelsberg (1986), Garcia and Gould (1987) and Zhu and Karee (1988) for some early empirical studies, and Basak and Mehra (2001) for a more recent investigation. Some people blamed the technique for the equity market crash on October 19, 1987. 7. According to an estimate in Bridgewater Associates (no date (b)), the expected turnover in spot contracts for dynamic replication of a one-year option is about 200%. 8. You might ask why there could be uncertainty here: from the option-replication example earlier, as long as someone follows the DIY guidebook religiously, the end result should be guaranteed irrespective of what actually happens afterwards. One main reason is that the option premium is based on the forecast of future spot volatility and the forecast can turn out to be wrong. 9. See also Bridgewater’s research note “Currency observations: the dynamic hedging enigma”. 10. Three per cent maximum loss would work only in spot-on-forward terms. 11. Conrad (1989), Detemple and Jorion (1990) and Sorescu (2000) used US stock data and found a significant price effect on the underlying asset as a result of the introduction of stock options. Urrutia and Vu (2001) showed that the return and volatility may increase or decrease after the introduction of stock options. 12. In the equity market, Dumas, Fleming and Whaley (1998) compared the out-of-sample performance of a wide range of deterministic volatility function models and found their performance unsatisfactory. Buraschi and Jackwerth (2001) found that returns for both in- and outof-the-money options are needed for spanning. In the fixed-income market Collin-Dufresne and Goldstein (2001) found that changes in the term structure of swap rates have very limited explanatory power for returns on ATM straddles. Jagannathan, Kaplin and Sun (2001) and Longstaff, Santa-Clara and Schwartz (2001) have provided similar evidence. We have not seen any work on whether the spot currency market spans the option market. Practical observations on how the currency options market works lend some support to the view that currency options are a distinct type of security, not spanned by spot. For most market makers in currency options the trading book is run independently of the spot/forward book. Both position taking and hedging of the currency options trading book are usually done through the options market directly. 13. Of course, volatilities are generally not constant across different expiries (or maturities) either. The relationship of volatility and time is referred to as the term structure of implied volatility, analogous to the similar relationship between interest rates and time. There is a considerable literature on the forward bias present in the term structure of interest rates, where forward interest rates are said to generally overpredict the future term structure of interest rates, and investors could potentially exploit this. Campa and Chang (1998b), using daily volatility data with maturities from one month up to twelve months in four major currency pairs, showed that, unlike interest rates, the term structure of implied volatilities does predict the future shape of the implied volatility curve. Many banks provide short-term buy or sell recommendations for short-term volatilities via options or other instruments based on whether short-term volatilities are higher or lower than the long-term counterpart. As far as we know, the effectiveness of this is unproven.
370
CO2_with_index_2nd_amends.indd 370
23/05/2011 14:09
Options, Dynamic Hedging and Overlay
14. Well, almost. There are a few more subtle and often more technical issues, such as decay, front-end versus back-end volatility, and the spread between implied and actual volatilities. 15. Some may argue that most funds are restricted in what they can do. For example, many cannot sell options and therefore could not sell implied volatility even if they wanted to. However, there are instruments that will allow clients to sell implied volatility by buying options. 16. Standard deviation is used here only as an indicative measure of the dispersion of the implied over actual volatility spread, as there is no evidence to show that such spreads are normally distributed. 17. The short-term forecasting power of Garch and implied volatility can be significantly improved if we use intra-day returns. See Andersen and Bollerslev (1998). 18. See Mayhew (1995) and Figlewski (1997) for surveys. 19. See Canina and Figlewski (1993) and Christensen and Prabhala (1998) on equity-implied volatilities and Neuhaus (1995) for fixed income. 20. Muralidhar and Neelakandan (2001). 21. Gamma refers to the fact that replication positions are sensitive to short-term spot movements. 22. Benzoni, Collin-Dufresne and Goldstein (2005) try to rationalise the persistent “expensiveness” of OTM S&P 500 put options after the 1987 market crash. 23. A nice remark the author heard when someone tried to impress a mathematically savvy client was “Gamma so-and-so to death!”
371
CO2_with_index_2nd_amends.indd 371
23/05/2011 14:09
CO2_with_index_2nd_amends.indd 372
23/05/2011 14:09
9
Emerging-Market Currencies
9.1 BACKGROUND It is generally accepted that both emerging-market stocks and bonds represent distinct asset classes, particularly for long-term investors.1 They are characterised by their very high volatility as well as low correlation with the asset types available in developed economies. The growing importance of the emerging market is underpinned by its demographic trend (population increase) and superior longer-term growth prospects. Many financial advisers and consultants recommend that investors allocate up to 20% of their portfolio to emerging markets, either directly or indirectly via some alternative asset classes, roughly in line with the gross-domestic-product weight of emerging economies relative to the global total, about 10% higher than the recommended figure a decade ago. The global financial crisis and its aftermath have reinforced the perceived upward trend of emerging economies as a whole going forward and various emerging-market countries have been fighting against the tidal waves of hot money into their economies. In a Global Financial Stability Report (2010) issued by IMF, the talk was to manage the “upside risk” in emerging markets. Emerging economies looked rosy during the first decade of the 21st Century, but what about the emerging-market currencies? Since the early 2000s research papers have looked into the erstwhile neglected subject of currency risk management under the umbrella of emerging-market investment.2 However, both in terms of quantity and quality, they are yet to match those done on developed curren373
CO2_with_index_2nd_amends.indd 373
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
cies. For one thing, market conditions for emerging-market currencies are very fluid, which makes statistical analysis difficult and, when it is done, less relevant to prevalent market conditions. Even when the price data is available and stable over a period of time, the conclusions of such studies should be viewed with caution. For example, bid–offer spreads in emerging markets are often several magnitudes higher than their developed-market counterparts, and failing to take this explicitly into consideration will hamper the usefulness of any statistical investigation. Also, we have to distinguish between returns in local currency, US-dollar-denominated returns and US-dollar-hedged returns (often with offshore non-deliverable forward contracts), as the impact of these is even more pronounced than for mature currencies. Analysis using US-dollar-denominated returns to study the impact of currency is almost pointless. In this short chapter we look at the special features of emergingmarket currencies and their implications for the hedging debate. We review the three main trading strategies relevant for EM currencies: carry, trend and value, also taking a look at implied–actual volatility spreads in emerging-market currencies and consider whether they can be used for trading. We finish by proposing a pragmatic and dynamic approach to dealing with emerging-market currencies. Appendix 9A contains a list of emerging-market currencies categorised by their current convertibility. 9.2 THE HEDGING DEBATE IN EM CURRENCIES 9.2.1 Special characteristics of emerging-market currencies3 Here is a list of key features of emerging-market currencies (DeRosa, 1999): o periodic devaluation due to domestic economic problems or global contagion; o high interest rates; o onshore controls; o frequent central bank intervention; and o high bid–offer spreads. Those features have more or less remained the same; however, people’s perception of how risky those features really are has changed over recent years. 374
CO2_with_index_2nd_amends.indd 374
23/05/2011 14:09
Emerging-Market Currencies
Hard (2005) summarised the pros and cons of adding emergingmarket currencies into a currency programme. The plus factors included a wider opportunity set, diversification, risk premiums in emerging-market currencies, and extra global insight gained. The minus factors included difficulty in establishing competitive advantage, dilution to existing investment process, disproportionate resource requirement, the intertwining nature of emerging-market currency returns with local interest rates and, of course, high transaction cost. 9.2.2 More recent interest in EM currencies The emerging-market economies as a whole came out of the global financial crisis in much better shape compared with developed economies, and many people remain bullish on their long-term growth prospects. And, over longer horizon, higher economic growth is typically associated with the appreciation of relevant currencies. In the meantime, the perceived riskiness of emerging-market currencies has also declined. At the height of the crisis, around October 2008, JP Morgan EM foreign-exchange volatility index (EM-VXY index)4 shot up to over 30%, about 10 percentage points higher than the corresponding G7 index (VXY index). Two years later, at the end of 2010,VXY stood at 12.5%, but EM-VXY was at 11.5%. Figure 9.1 The currency and yield components of EM local currency debt returns 25% 20% 15% 10% 5% 0% -5% -10% -15%
FX Bond Yield Total 2003
2004
2005
2006
2007
2008
2009
2010
Source: Bloomberg, BNP Paribas Investment Partners
375
CO2_with_index_2nd_amends.indd 375
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
For some people, investing in emerging-market currencies is a useful alternative to investing in emerging-market local debt, because a large portion of the return to EM local debt comes from the appreciation of EM currencies. Figure 9.1 shows the breakdown of EM local-currency debt returns over recent years and it is clear that large proportion of the returns was from the currencies. Pareto Investment Management, a currency-overlay specialist, argued in a research paper (2009) that, since currency dominates EM local bond returns, it would not make sense to follow debt benchmarks that are based on the relative size of the local debt market rather than the expected returns from EM currencies. The paper added that EM currency portfolios would have the flexibility to take both long and short positions in the EM currencies versus the US dollar, or against each other, and are therefore better suited to take advantage of the diverging growth differentials between the emerging-market economies and reduce the total portfolio volatility vis-à-vis that of the long-only EM bond portfolios. Using both the unhedged and hedged JP Morgan emerging-market government bond indexes between 2002 and 2008, the report noted that the former has an annualised return of 12.78% with volatility of 11.05%, better than the latter with a return of 5.32% against volatility of 4.26% on simple return terms, without risk adjustment. The view was eched in Dissaux (2010), which also pointed out that access to and the quality of the macroeconomic data from these EM countries are much better than before, due to transparency drives carried out by the governments; also, the risk of capital control in emerging markets, though still not to be ruled out, can be better anticipated and managed than before. In addition, the liquidity situation in all forms of financial instruments linked to EM currencies has improved significantly, evidenced both by the latest BIS foreign-exchange survey that we have reviewed in Chapter 1 and further documented by Saxena and Villar (2008). 9.2.3 Should you hedge emerging currency exposures? It is not unusual to hear stories like the following from investors. They decided to hedge all their currency exposures in the emerging market but, because things were quiet for a time, they switched off the hedging programme due to the high rolling cost. When devalu376
CO2_with_index_2nd_amends.indd 376
23/05/2011 14:09
Emerging-Market Currencies
ation happened – around which time it was too late or expensive to reinstate the hedges – the underlying investment tumbled in value, often both in local currency and in translated terms. In effect, the investor pays twice. With its high risk, it would seem logical for an investor to hedge exposure to an emerging-market currency completely. However, there are good practical reasons not to: o hedging instruments are unavailable or too illiquid and the cost of hedging is high; o the allocation to EM currencies is usually small; o many emerging-market currencies are pegged against the US dollar or a basket of mature currencies, which reduces the need to hedge;5 and o very few specialist emerging-market currency managers and products are available. There are other, more important, factors against hedging risks in emerging currencies. First, from a return perspective, emerging-market currencies tend to exhibit consistent positive carry as well as long-term appreciation, and hedging will turn these advantages into disadvantages. Second, from a risk perspective, if an investor holds a diverse mix of EM currency exposures, then as a portfolio the currency risk is usually substantially lower, even if we take the tail risk of large devaluations into consideration. Generally speaking, the correlation between emerging-market currencies in the same region is relatively high, especially during a crisis, but the correlation between emerging-market currencies in different regions is low and often negative, hence a portfolio of EM currencies tends to exhibit a much lower risk level than that of its components. Eaker, Grant and Woodard (2000) investigated this point in detail with a portfolio of nine emerging-market equity investments between 1975 and 1997. They found that, for US investors, diversification virtually eliminates the impact of currencies and that there may be a substantial diversification benefit if emerging-market currency risks are considered in a basket. They also investigated the highest and lowest equity returns in US dollar terms every year over the same period, 377
CO2_with_index_2nd_amends.indd 377
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
breaking these down into currency returns and equity returns in the local currency. They found that, out of 44 of the highest and lowest dollar-denominated equity annual returns, 29 were associated with depreciation of the local currency against the dollar by more than 10%.6 In other words, big swings in emerging-market equity returns often coincide with a large depreciation of the local currency. However, not all of the 29 cases could be described as double whammies (where the investor suffers losses in both local equities and currency in the same year), as the authors found that in 12 of these the dollar-denominated returns were in fact the highest of that year. In other words, when an emerging-market currency devalues, the underlying investment may still produce a good return in the investor’s base currency. The point on diversification is also illustrated in Burnside, Eichenbaum and Rebelo (2007): in their EM currency carry trade portfolio, some currencies could lose more than 50% in a single week, but the losses on the portfolio are somewhat limited. A similar study by Acar and Middleton (2003) looked at the hedging issue related to equity investments into four countries, Poland, Turkey, Mexico and South Africa, between March 1998 and October 2002, and they also failed to find any conclusive evidence in support of currency hedging. More recently, Normand et al (2010) noted that EM currencies have been consistently adding to investment returns by providing both carry and appreciation across different regional blocks, and, therefore, EM currencies should be treated as an exception – as opposed to their developed counterparts, which are usually mean-reverting over time – when the question of hedging is asked. A counterpoint was provided in Muralidhar (2001), which used data from December 1993 and November 2000 and contended that unhedged EM currency distracted value for equity investors in those countries. The conclusion was, however, based on the local currency equity returns versus the (unhedged) returns in US dollars without taking historical hedging costs (interest rate differentials) into consideration.7 Compared with the lengthy survey on the hedging debate for development currencies in Chapter 1, Section 1.6.4, it is easy to see that we are on shaky ground for EM currencies. On balance, we believe that systematic hedging of EM currencies is probably unwarranted in most cases. 378
CO2_with_index_2nd_amends.indd 378
23/05/2011 14:09
Emerging-Market Currencies
9.2.4 Hedging selectively Some consultants or investment advisors maintain that investors should hedge emerging-market currencies only before an impending currency crisis. It is, of course, very sensible advice. Maybe too sensible. In order to do this, you would need not only to predict the timing of the currency crisis but also to do it before most other market participants in order to secure favourable pricing terms on hedging contracts. Unfortunately, neither is easy. The topic of EM currency crises has been investigated extensively in the past. There are two main explanations for the devaluation of emerging-market currencies.8 The first focuses on economic fundamentals and argues that errors in government fiscal and monetary policies lead to speculative attacks and subsequent devaluation when the means of defence are exhausted. According to this explanation, devaluation is partially predictable from stress signals manifest in, say, growing budget deficits, fast monetary growth, rising wages and prices, or overvaluation of real exchange rates. The second explanation points to self-fulfilling crises caused by speculative attacks motivated by what Keynes called “animal spirits”.9 By analysing 96 currency crises between 1970 and 2002, Kaminsky (2006) suggested that the weakness in fundamentals has been a more important cause of crises, though the weakness could manifest itself in many different forms. The prediction methods of EM currency crises, usually known as the “early-warning system” or “EWS” among researchers, come in many flavours, mainly due to large numbers of potential EWS indicators: for example, according to Chamon, Manasse and Prati (2007) of the IMF, measures of real exchange-rate misalignment; terms of trade shocks; international reserves; external and public debt; monetary and fiscal policy; balance-sheet mismatches (currency and maturity) in the corporate, banking and government sectors; political uncertainty; global cyclical and financial conditions; and general market sentiment. There would be several subindicators within each category, but trying to sift through them and identify the ones that truly matter is not an easy task. The three authors used a statistical technique called “binary classification trees” to navigate mounds of prices, macroeconomic data and forecasts. Their EWS did well in in-sample testing, with an accuracy rate of 379
CO2_with_index_2nd_amends.indd 379
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
88%. It did not foretell all the crises in out-of-sample testing but the result seemed good enough to be applied in real portfolios. For example, del Vecchio (2006) of JP Morgan Investment suggested a very practical “signal extraction approach” to forecasting EM currency crises – similar to the binary classification of trees – using indicators on current and capital accounts, and on real and financial sectors.10 Overall, existing empirical studies11 seem to indicate that currency devaluation may to some degree be predictable and selective hedging probably makes sense for some investors. Many currencyoverlay managers now offer this service to their clients. Even if you did have some way of seeing a devaluation coming, you would have to be far enough ahead of the market to put a hedge in place but not so far that the cumulative cost of rolling the hedge outweighed the eventual benefit of hedging when the crisis materialised. Markets often react quickly when the whiff of crisis is in the air, and both forwards and options volatilities shoot up, which pushes up the cost of hedging. 9.2.5 Hedging issues for investors based in emerging-market currencies Most of the books on currency risk management, this title included, assume the vantage point that the investor’s base currency is one of the developed currencies and their risky currency exposures are either in other developed currencies or in emerging-market currencies. With the rapid expansion of institutional investors (including central banks and sovereign wealth funds) based in emerging-market economies and the increasing allocation they hold in foreign currencies, this situation is clearly unsatisfactory. In many ways, the hedging problem they have to deal with is even trickier than that for investors based in developed economies: their base currencies are on an upward trend; their hedging costs are high, due to high domestic interest rates and high transaction costs; hedging instruments are unavailable or not traded with liquidity; they are forced to hold a lot of US dollar exposures for reasons not linked to seeking return or reducing risk but they are blamed for the markto-market losses suffered on these holdings; and the list goes on. Some of the methods and solutions – proxy hedging, basket hedging, for example – discussed in this book can be applied to these 380
CO2_with_index_2nd_amends.indd 380
23/05/2011 14:09
Emerging-Market Currencies
situations, but many others need reworking. For instance, rather than think about hedging back to base currency, an investor based in an emerging-market currency could use a currency overlay to spread the currency among many different currencies in order to reduce risk. Many such issues have only began to be investigated by researchers and practitioners after the global financial crisis. 9.3 CARRY, TREND AND VALUE IN EM CURRENCIES 9.3.1 Carry in EM currencies We have reviewed the large body of literature on the topic of forward bias in Chapter 4, Section 4.5.1. Most works on the topic consider the currencies of developed countries. However, more and more studies have since systematically investigated forward bias in emerging-market currencies. Francis, Hasan and Hunter (2002) have proposed that forward bias in emerging markets is systematic in nature and more likely to be larger and more persistent than in developed markets. Flood and Rose (2002) found that forward bias “improved” (ie, became less pronounced) in the 1990s over previous periods but that (1) it still exists across different countries and (2) there is no significant difference between developed and developing countries. Bansal and Dahlquist (2000), however, found that forward bias is more frequently observed in developed economies and less frequently in emerging-market currencies, and their view was shared by Frankel and Poonawalaa (2006). Bansal and Dahlquist (2000) found that including emergingmarket currencies in portfolios containing investments in the developed markets improves the portfolios’ Sharpe Ratios. Malliaropulos (1997) concluded that the expected excess returns of foreign currency deposits are less volatile than those of equities and that the addition of US dollar deposits to an international equity portfolio can provide additional diversification benefits to non-US investors. Bridgewater Associates (no date (c)), using 15 emerging-market currencies over a period longer than 20 years (1972–96), have estimated that excess returns on emerging-market currencies amount to about 2–3% per annum over a long horizon. They also comment that expected volatility in any single currency will be around 10– 15%. However, by combining these as a portfolio, the risk can be reduced to around 5%. 381
CO2_with_index_2nd_amends.indd 381
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
As noted earlier, many previous studies on EM currencies suffered from data problems: incomplete datasets and small number of currencies, non-tradable prices (for example, using onshore prices), incorrect transact costs, and so on. Newer attempts generally consider those issues much more carefully. More recently, in an oft-quoted paper, Burnside, Eichenbaum and Rebelo (2007) compared the carry trade returns from both developed currencies and EM currencies. By incorporating the relevant bid–offer spreads and looking at a large number of EM currencies, they concluded that including EM currencies would significantly improve the risk-adjusted returns. Gilmore and Hayashi (2008) further observed that, since the late 1990s, the broad basket of emerging-market currencies has provided significant equity-like excess returns against a number of major market currencies, but with low volatility, and a large part of that return came from carry – the remainder from the “currency surprise”, ie, additional spot appreciation. Instead of nominal interest-rate differentials, de Zwarty et al (2008) used the real interest-rate differential of 21 emerging markets between 1997 and 2007 and obtained statistically significant risk-adjusted returns. The subject of carry trade in EM currencies is currently an actively researched area. Many new empirical studies seem to confirm that that adding emerging-market currencies to an investor’s portfolio significantly improves the portfolio’s performance – in other words, currency risk in the emerging markets, unlike in developed markets, should not be systematically hedged. Carry trades are routinely used by many players in the currency markets. However, the mechanics of the trade mean that investors accumulate profit in small measures over time but are exposed to the risk of periodical devaluation. Also, it has been noted that carry trades with stop-loss orders may lead to episodes of sharp exchange-rate movements and loss of liquidity in the trading of emerging-market currencies.12 There are several beta indexes tracking the carry performance of either emerging-market currencies alone or together with G10 currencies. For example, Deutsche Bank’s Global Harvest Index contains five highest-yielding currencies against five lowest from G10 and EM.13 It has yielded impressive returns since inception. For comparison, we also plotted the G10 Harvest Index from the same 382
CO2_with_index_2nd_amends.indd 382
23/05/2011 14:09
Emerging-Market Currencies
source. The global index includes USD, EUR, JPY, GBP, CHF, AUD, CAD, SEK, NOK, KRW, MXN, NZD, SGD, ZAR, PLN, TWD, BRL, CZK, HUF and TRY, a much larger universe than the G10. Figure 9.2 Carry trading including and excluding emerging-market currencies 240
DB G10 Currency Harvest ER DB Global Currency Harvest ER
220 200 180 160 140 120 100 80 2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Source: Deutsche Bank, available at www.dbiq.com
9.3.2 Trend-following in EM currencies Studies on the effectiveness of trend-following strategies in EM currencies are not common. EM currencies tend to trend up on a long-term basis, interrupted by an occasional large setback; their paths tend to be heavily influenced by intervention effects of their respective central banks; their high transaction costs do not favour frequent trading. All these would mean that, if you want to follow trend in EM currencies, you follow the medium-to-long-term trend, you do not trade frequently and you try to reverse the positions before any crisis. In other words, trend-following in EM currencies is like carry trade. However, there are a few emerging-market currencies that have been traded liquidly for a long while and their trajectories less controlled by their central banks. On those currencies, a more dynamic trend approach might have worked. For example, in an early study, Martin (2001) applied moving-average rules on 12 EM currencies from 1993 and 1995 for 12 emerging markets and 383
CO2_with_index_2nd_amends.indd 383
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
reported significant out-of-sample excess returns. Lee, Gleason and Mathur (2001) used both moving-average and channel-breakout rules on Latin American currencies and found that those rules worked for some currencies but not for others. Martin’s study used an assumption of a fixed transaction cost, whereas Lee et al’s study did not take interest-rate differentials into considerations. Pojarliev (2005) employed two simple trend-following rules: one looks at the return from the previous month and takes long position if the return is positive and vice versa; the other compares the current spot exchange rate versus the three-month moving average and goes long if the current spot is above the moving average and short otherwise. He reported an information ratio of 1.48 between 1999 and 2004. More recently, a similar information ratio was reported by Secmen (2009) in a research note by Citibank, using data from 1997 to 2009. De Zwarty et al (2008) also tested various combinations of moving-average crossover rules and channelbreakout rules and stated similar profitability levels on combined trading portfolios. We believe that, with more and more EM currencies becoming free-floating and showing healthy two-way moves, there ought to be more opportunities for deploying these trading techniques previously applied only to developed currencies. However, a word of caution here. The reported impressive historical returns could all potentially suffer from the data-snooping problems mentioned in Chapter 4: that researchers and practitioners, knowingly or not, selected the rules that fit the historical data and thus show the best trading results. In recent years, some more powerful statistical tools have been developed to provide a reality check on whether the reported profits represent genuine trading opportunities or are merely there by chance. Qi and Wu (2006) checked more than 2,000 moving-average rules on EM currencies and found that correcting for the data-snooping problem would have reduced the historical profitability, but they would have remained profitable. They also reported that these rules have become less lucrative over time. Kuang, Schröder and Wang (2008) studied close to 26,000 rules and concluded that the profit from momentum trading in EM currencies is all illusory.
384
CO2_with_index_2nd_amends.indd 384
23/05/2011 14:09
Emerging-Market Currencies
9.3.3 Value-based trading in EM currencies On PPP-based measures, EM currencies are generally undervalued and the only position we ought take is to go long. De Zwarty et al (2008) also tested an alternative rule based on the differentials in GDP growth rates between EM countries and the US and found that the returns from this simple strategy are better than those from carry-trade strategies. All three types of strategy mentioned in this chapter were tested in de Zwarty et al and the authors examined the performance from combined portfolios and printed Sharpe Ratios close to 1.5. One alternative to PPP was the Fundamental Equilibrium Exchange Rate (FEER), promoted and regularly estimated by the Peterson Institute for International Economics (Williamson 1983; Cline 2008). FEER represents the hypothetical equilibrium exchange rates that “generate a current account surplus or deficit that matched the country’s underlying capital flow over the cycle, assuming that the country is pursuing internal balance as best as it can and that it is not restricting trade for balance-of-payments reasons”. FEER estimates are usually much more closer to the market exchange rates, and large gaps between them could be used as trading signals for value-based trading in EM currencies (Dissaux 2010). 9.4 USE OPTIONS TO TRADE EM CURRENCIES 9.4.1 Overview Research on the behaviour of implied volatility in emerging-market currencies is scarce, mainly due to lack of data. In practice we usually observe that implied volatility is itself more volatile and the implied–actual volatility spread is more pronounced than in developed markets. For the investor or currency manager information embedded in the implied volatility is important because: o implied volatility can be exploited directly as a source of additional return; o implied volatility, rather than actual volatility, should be used to estimate hedging costs (so that an informed decision can be taken); and o implied volatility can be more informative of potential future movements than actual volatility, especially for pegged emerging-market currencies, and this information can be incorporated in the risk management process. 385
CO2_with_index_2nd_amends.indd 385
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
9.4.2 Selling volatility If implied volatility is systematically higher than actual volatility, an investor can sell an option to exploit the discrepancy just as in mature currencies. Generally speaking, this type of strategy where investors accumulate small returns and suffer large losses, albeit only occasionally, is probably not suitable for those with a low risk appetite. 9.4.3 Buying call options on emerging-market currencies to capture carry Some trading strategies use options instead of forwards for carry trade: carry with some form of protection. Consistently buying options on emerging-market currencies would probably result in a loss, hence this type of strategies would need to use some optimisation techniques to (1) buy options selectively; (2) weight the relative positions according to their risk–reward potentials and how they interact with one another; (3) use some risk indexes to unwind the positions. We have seen some banks selling this type of product. Usually, they perform much better during backtest periods than in live trading. 9.4.4 Using options to forecast spot Is there potentially additional information on future spot movements embedded in option skews – ie, when a call option on a currency is more expensive than the put option of the same delta, does that signal that the currency may appreciate? It would be interesting to see whether the implied volatility of emerging-market currencies can be used to forecast spot movements. The only research in this area that we are aware of is by Campa, Chang and Refalo (1999). Using currency option data from the local Brazilian exchange to test market perception of the credibility of the crawling peg regime, they found evidence that information embedded in option prices could indeed provide a better forecast of periodic realignments of the exchangerate bands than macroeconomic or interest rate-based indicators. 9.5 OTHER RELATED ISSUES 9.5.1 Cross-correlation We have seen that investors or fund managers sometimes use proxies to hedge emerging-market currencies. For example, they may treat exposure to the Korean won or Singapore dollar as an exposure to, 386
CO2_with_index_2nd_amends.indd 386
23/05/2011 14:09
Emerging-Market Currencies
respectively, the Japanese yen or US dollar and hedge accordingly. Hedging the yen or dollar obviously attracts much lower direct trading costs than trading in the won or Singapore dollar. The underlying assumption of this practice is a stable and high correlation between the Korean won or Singapore dollar and the Japanese yen or US dollar. In ordinary times such assumptions will usually hold, as they are often a direct consequence of the relevant central banks’ foreignexchange policies. However, during financial turbulence the policy could be abandoned and the correlation break down, resulting in a potential for big mismatches between the hedge and the underlying. Apart from ad hoc research papers from investment banks and some specialist fund managers, the only systematic study of this approach we know is that by Aggarwal and Demaskey (1997). They looked at the hedging performance achieved by using the Japanese yen, sterling, the Canadian dollar, the Deutschmark and the Swiss franc to hedge currency risks in the Hong Kong dollar, Korean won, Singapore dollar, Taiwan dollar, Indonesian rupiah, Philippine peso and Thai baht over the period 1983–92, using Sharpe Ratio as the performance measure. They found that crosshedging generally did improve the performance for US-based investors and that the yen was the best proxy in most cases apart from the Hong Kong, Singapore and Taiwan dollars, currencies that move closely with the US dollar. However, it would be very difficult to apply their approach in practice as their minimumvariance hedge ratio varied from 900% to 3,400%. Another related issue is the cross-correlation between emergingmarket currencies themselves. Low correlation between these currencies indicates a high potential for diversification, and vice versa. We are not aware of a direct attempt to quantify such cross-correlation, but the IMF’s Emerging Market Bond Cross Correlation Index can be used as a close proxy.14 9.5.2 Basket peg At the end of December 1996, at least 30 countries were using a basketpegging approach according to data compiled by the IMF. Some other currencies, notably the Singapore dollar and South African rand, may also include a pegging element. Some market participants have historically been using carry trades to explore the interest-rate differentials between the basket currencies and the emerging-market currency in 387
CO2_with_index_2nd_amends.indd 387
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
question.15 A study by Christoffersen and Giorgianni (1999) highlights the danger of such an approach. They developed an ex ante method to estimate time-varying weights and ex ante risk measures for a proxy hedging approach. The motivation was to analyse the potential profit/ loss from carrying trades where the monetary authority of the underlying currency follows a policy of linking its currency to a basket of hard currencies but with an undeclared, and potentially time-varying, weight. They found that traditional rolling regression methods may significantly underestimate the risk involved. Taking the Thai baht as an example, when the basket peg was abandoned on July 2, 1997, the baht dropped 9.8% on that day. Using the estimated daily profit figure of 0.021% per trading day, it would take 445 trading days of average profit to make up the loss experienced on that particular day – not to mention the trading restrictions introduced after the devaluation. For comparison, in the rolling model, which is what most people use to estimate the risk in carry trades, the crash represents a 29-standarddeviation event – a near impossibility. Applying Christoffersen and Giorgianni’s approach, it is only a 10-standard-deviation event – still quite unlikely but not impossible. 9.5.3 Indexes of risk and risk appetite A number of financial institutions have developed indexes that try to capture the risk of investing in certain markets and/or the risk appetite of investors. This kind of risk index can be used in conjunction with any foreign-exchange trading models discussed in Chapter 4, in particular carry trade models. It is also worth pointing out that risk or risk-appetite indexes can be applied on both developedmarket currencies and emerging-market currencies. The indexes of risk appetite are often the more subjective, some relying on investors’ speculative positions, others based on survey data. One approach proposed by Kumar and Persaud (2001) in an IMF paper is based on a special type of correlation between the excess returns of currencies and the volatility of those returns. The rationale of such an approach is that a high correlation indicates that the return is commensurate with the associated risk, which in turn signifies a rising risk appetite. The authors suggest that, when their risk-appetite index goes up, high-interest-rate currencies (often emerging-market currencies) will probably outperform. 388
CO2_with_index_2nd_amends.indd 388
23/05/2011 14:09
Emerging-Market Currencies
The risk indexes are usually a statistical summary of a selection of financial-market variables: for example, liquidity, actual volatility, implied volatility, option skews, credit spread and correlations between some variables. A well-known index in this camp is the Liquidity, Credit and Volatility Index (LCVI) computed by JP Morgan Chase, the US investment bank.16 Many other banks now have their versions of risk indexes, and so do some central banks and think tanks. The European Central Bank has a Global Hazard Index, which is frequently used by market participants to condition their positions in carry trades and/or in emerging-market currencies. UBS produces a daily index called the FX Risk Index+ (Figure 9.3), derived from seven factors: equity-market volatility (VIX); EURUSD FX option implied volatility; USDJPY FX option implied volatility; gold prices in EUR and USD; the difference in performance between the S&P financial and utility equity sectors; o the spread of corporate bonds over US Treasuries; and o the relationship between US bond and stock prices. o o o o o
Figure 9.3 UBS’s FX Risk Index + 1998–2010 4.0
UBS Risk Index +
3.0
2.0
1.0
0.0
-1.0
-2.0 1998
2000
2002
2004
2006
2008
Source: UBS AG
389
CO2_with_index_2nd_amends.indd 389
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
The UBS FX Risk Index+ aims to identify periods of risk-averse and risk-seeking behaviour displayed by investors. Each component is represented by its Z-score over the past year. The final index is given by average value of these components. By adjusting our position in sympathy with the state of the Risk Index+ we aim to avoid prolonged draw-down periods 9.6 SUMMARY On balance, the evidence we have reviewed in this chapter suggests that a blanket approach to emerging-market currency exposures may not be appropriate, especially in view of the high hedging cost in both forwards and options markets. We would recommend a more pragmatic and dynamic programme to deal with specific problems and changing market conditions. CalPERS, the California Public Employees’ Retirement System, together with consultant firm Wilshire Associates, has developed an approach to investing in emerging markets using a scoring system on a host of country- and market-related factors, including political stability, market regulation and volatility, liquidity, legal systems and transparency. The same approach can be adopted for currency hedging, incorporating signals in budget deficits, monetary growth, wage and price inflation, real exchange-rate overvaluation, etc. One important factor in the decision-making process should be transaction costs. Estimates of these should be incorporated into the calculation of potential return enhancement and/or risk reduction.
1. Some recent research (eg, Garza-Gómez and Metghalchi 2006) has started to question this premise, citing the increasing level of cross-correlation between emerging-market asset returns and developed-market asset returns as evidence. Given the disparate nature of emerging markets and the varying degree of “emergedness” of each market, we will stay with the established the view here. 2. For general discussions of (and references for) emerging-market investment, see Errunza (1997) and Aggarwal (1994). Hauser, Marcus and Yaari’s (1994) early empirical study concludes that there is no benefit. Bridgewater Associates has produced two very informative research papers on emerging-market currencies. Both are available from its website at www. bridgewaterassociates.com. 3. The International Monetary Fund’s semiannual Global Financial Stability Report has useful information on market activities in emerging currencies. Available from www.imf.org/external/pubs/ft/gfsr/.
390
CO2_with_index_2nd_amends.indd 390
23/05/2011 14:09
Emerging-Market Currencies
4. The JP Morgan VXY and EM-VXY indexes follow aggregate volatility in currencies through a turnover-weighted index of G7 and emerging-market volatility, based on three-month atthe-money forward options. The indexes are designed to allow investors to measure aggregate risk premiums in currency markets, calibrate trading strategies and express views on volatility as an asset class. 5. Some emerging-market debt is US dollar-based – for example, Brady bonds and other dollar-denominated securities. These securities have no direct currency exposure but full credit exposure. 6. Note that in Eaker, Grant and Woodard (2000) currency returns are calculated by spot-onspot changes and are therefore not necessarily hedgeable. 7. It may seem a bit unfair to criticise the article on the point of hedging cost (interest rate differential) but we mention it here since we also comment on Eaker, Grant and Woodard (2000) – see previous endnote. Eaker et al argued against hedging where the hedging cost is not paid directly but Muralidhar argued for; the issue of hedging cost is therefore more important for the latter. 8. See the comprehensive surveys by Esquivel and Larrain (2000) and Bekaert and Harvey (2003) for further references. 9. Keynes (1936) and Howitt and McAffee (1992). 10. See also Ghosh and Ghosh (2003), Frankel and Wei (2004) and Kaminsky (2006) for further references. 11. Berg, Borensztein and Pattillo (2005) contains a survey of similar EWS. 12. European Central Bank (2003), p. 21. 13. Available at https://index.db.com/index/balanced_currency_harvest_usd. 14. The index appears in the IMF’s Global Financial Stability Report – see note 3. 15. For historical studies, see Horngren and Vredin (1989), Klein (1989) and Pikkarainen (1991). 16. The LCVI also contains a risk-appetite element.
391
CO2_with_index_2nd_amends.indd 391
23/05/2011 14:09
CO2_with_index_2nd_amends.indd 392
23/05/2011 14:09
10
Liquidity, Counterparty Risk and E Commerce
10.1 INTRODUCTION In this chapter we bundle together three short and seemingly unrelated topics: foreign-exchange liquidity, counterparty risk and foreign-exchange e-commerce. The first, foreign-exchange liquidity, is widely talked about but not necessarily well studied – not surprising, given the predominantly over-the-counter nature of the foreign-exchange market and, consequently, the lack of data we would need to conduct such a study. The topic of foreign-exchange liquidity is often mentioned by practitioners in their jargon-peppered talks: piping liquidity this way, pooling liquidity that way. We first present a short overview of foreign-exchange liquidity, including both academic findings and practical observations, with the aim of providing some background information for someone without prior theoretical or practical knowledge in this area. Then we discuss the changes to foreign-exchange liquidity since 2000 and the main drivers behind the change. The issue of counterparty risk was pushed into centre stage at the onset of the global financial crisis. Previously, the issue of credit risk had not been considered as a serious problem for currency-overlay programmes, because only over-the-counter forwards, swaps and options with very short maturities are used in those programmes. It was hard to foresee that one of the counterparty banks, usually household names with sprawling global networks, could go under during the life of a forward contract, mostly one to three months. 393
CO2_with_index_2nd_amends.indd 393
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
And we all know now – painfully for some – that they could, and did. We discuss the methods that investors and currency-overlay managers use to mitigate the counterparty risk and how these affect pricing and market liquidity. Since 2000 the topic of e-commerce has gained a firm place in foreign-exchange discussions (even though electronic dealing has been around since 1990 or even earlier) and it is fundamentally changing the operation of the foreign-exchange market, including the provision of and access to foreign-exchange liquidity. Hence we opt to combine these three topics in one chapter. Since late 1990s, innovations in foreign-exchange e-commerce have changed not only the way foreign-exchange transactions are done but also many other aspects of the life cycle of such transactions: matching confirmation, settlement, portfolio management and so forth. The benefits of online transactions, via either a multibank portal or a single-bank system, and STP are easy to see, and we will use FXall and UBS, respectively, as examples to illustrate them. We are witnessing significant efficiency improvement when online innovations are combined with the currency-overlay process – something we predicted in the previous edition and is now materialising. 10.2 FOREIGN-EXCHANGE LIQUIDITY: THEORY AND REALITY 10.2.1 Theory of foreign-exchange liquidity Foreign exchange is the largest financial market, with massive trading volumes – far larger than supported by the underlying trade flows between the countries. There are many theories on this socalled excess volume, mostly involving “noisy traders”: large number of trades react to new information flows, interpret them differently and trade differently. Lyons’s (1997) “hot potato” theory provides an alternative explanation. It works as follows: when a client executes a large trade through a marker maker, the latter eliminates all or part of the exposure to minimise the risk of adverse price movements. They can call out on the interbank market and offload their exposure by hitting on other dealers’ quotes.1 The dealers hit by such positions may then wish to call out on other dealers to pass the exposure on. The original trade is thus passed around like a “hot potato”, multiplying itself several times en route. Lyons’s theory fits well with two characteristics of the foreign394
CO2_with_index_2nd_amends.indd 394
23/05/2011 14:09
Liquidity, Counterparty Risk and E Commerce
exchange market: its dealer structure and its lack of trade reporting. If the theory is correct, better mechanisms of price dissemination could fundamentally reduce the volume, whereas the impact might not be so drastic if the noisy-trader theory were true. In fact, you could argue that the volume from noisy traders might actually increase if price transparency increases. The advent of electronic trading/brokering, eg, EBS and Reuters’ D2000-2 systems, has fundamentally changed the price discovery process, and the online multibank portals are pushing the boundary even further in that these B2C (business-to-consumer) tools, such as FXall, are increasing the price transparency for customers. We had indeed seen a large reduction in foreign-exchange volume between 1998 and 2001, as indicated in the triennial BIS surveys; but the volume quickly rebounded to an astonishing level in the latest BIS survey in 2010 (see Figure 1.2). The latest evidence shows that the “hot potato” trading activity is dwindling but the army of noisy traders is swelling. King and Rime (2010) from BIS reported that due to the increasing flows through banks’ electronic dealing platforms, banks have been able to match a significantly larger portion of trades against each other and therefore capture the bid–offer spread. The top foreign-exchange banks reported to BIS’s triennial survey that, in 2007, they matched about 25% of all trades, and the ratio increased to 80% in the 2010 survey. The remaining 20% of trades are the “hot potatoes” placed to the market and in relative terms the volume is now much smaller than before. At the same time, King and Rime (2010) found evidence that three categories of traders probably explained the bulk of increasing foreign-exchange turnover over recent years. o High-frequency/algorithmic traders: Some estimate that they account for between 25% and 50% of total volume in foreignexchange spot and futures trading. o Smaller banks: As we explained in Chapter 1, many of them have become “customers” of large banks and focus on marketing to their clients. o Retail investors: King and Rime (2010) estimated that they now probably account for 8–10% of global FX spot turnover.
395
CO2_with_index_2nd_amends.indd 395
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
In addition, King and Mallo (2010) from BIS noted that 16% of all foreign-exchange transactions are conducted via prime broker from a survey done by the London Foreign Exchange Change Committee in April 2010. These trades are counted twice in BIS’s surveys. We believe this too is an important factor behind the dramatic growth in turnover. There are other types of foreign-exchange liquidity models that attempt to explain the dissemination of information and the role of private information in foreign-exchange trading. For example, Ito, Lyons and Melvin (1998) analysed data on US dollar/Japanese yen trading over a particular day when a certain trading restriction was lifted and found evidence consistent with the presence of significant private information in the foreign-exchange market. Covrig and Melvin (1998) arrived at a similar conclusion. They concluded that, at the intra-day level, private information is a very important factor for foreign-exchange trading, even though its importance diminishes markedly for longer time periods and/or lower-frequency data. Peiers (1997) observed trading activities of the Deutsche Bank in the US dollar/Deutschmark around periods of Bundesbank intervention and concluded that the Deutsche Bank is a price leader over those periods. He suggested that the bank received a significant portion of the intervention-related currency trades and therefore had a temporary information advantage. His conclusion is supported by Sapp (2002). De Jong et al (1999) repeated Peiers’s analysis but did not find evidence of price leadership. As we mentioned in Chapter 4, Section 4.6.5, the microstructure of the foreign-exchange market, where foreign-exchange liquidity is also studied, is the subject of active research. Interested readers can follow up Lyons (2002) and the references therein.2 10.2.2 Electronic dealing and its impact on liquidity The spread of electronic execution in the foreign-exchange markets during 2000s was transforming the financial landscape. According to Greenwich Associates, more than 50% of all foreign-exchange transactions are conducted electronically. If we include the transactions via electronic brokers, the majority of foreign-exchange transactions are done using electronic means. According to a BIS survey in 2010, 41.3% of all foreign-exchange transactions are done via electronic 396
CO2_with_index_2nd_amends.indd 396
23/05/2011 14:09
Liquidity, Counterparty Risk and E Commerce
methods, 24.3% interbank, 18.5% customer direct and 15.9% via voice broker. However, nearly all interbank trades are electronic, so are more than half of the customer trades. Hence close to 80% of all foreign-exchange transactions are done electronically, out of which the split between single-bank platform, multibank platform and electronic broking systems are roughly 20%, 20% and 40%. The prevalence of electronic trading and electronic broking reduces transaction costs and increases liquidity, and it has paved the way for the growth of algorithmic trading and high-frequency trading (King and Rime 2010). The pace of change is fast. In the 1992 survey, electronic brokers channelling only about 2% of total trades and multi- or singlebank platforms were nowhere to be seen. In the 1998 survey, close to half of the trades were done via electronic brokers. All the electronic trading or broking systems are known as electronic communication networks (ECNs). They include electronic broking systems such as EBS and Thomson Reuters Matching; multibank trading platforms such as FXall, FX Connect, FXCM, Hotspot FX and Currenex; and single-bank trading platforms such as UBS’s FXTrader, Deutsche Bank’s Autobahn, Barclay’s BARX and Citigroup’s Velocity. All three types of network experienced increase in volume over recent years. During and after the global financial crisis, single-bank platforms have grown at a higher pace. The banks usually give priority to their own trading platforms than multibank ones, especially during periods of market stress. Huang and Masulis’s (1999) study, using data for 1992–93, indicates that bid–offer spreads increase with foreign-exchange volatility and decrease with competition. However, the bid–offer spread in the spot foreign-exchange market had been relatively stable between 2005 and 2010 for all mature currency pairs, barring the one-off adjustment during the introduction of the euro. Goodhart, Ito and Payne (1996), in a study of trading activities via the Reuters D2000-2 electronic brokering system, observed that bid–offer spreads in the foreign-exchange market are determined by market convention rather than potential cost. More recently, Ding and Hiltrop (2010) found that the introduction of these electronic trading methods, especially EBS, has reduced the bid–offer spread in foreign-exchange trading significantly and thus improved market liquidity. They also found that the geographical variation in foreign397
CO2_with_index_2nd_amends.indd 397
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
exchange liquidity has been greatly reduced. The beneficial impact of electronic dealing methods is also surveyed in Barker (2007). Transaction costs spiked during the crisis, both in spot markets and swap markets; Figure 10.1 shows the three-month forward spreads in four currency pairs over the crisis period and a threemonth moving average of all four series. Many price takers complained that the banks were quick to raise the spreads during turmoil but slow to reduce them later, especially in view of the bumper foreign-exchange trading profits reported by banks after the crisis. There were rumours that some large banks, fewer in number during the crisis, formed secret cartels in certain currency pairs (for example, euro against British pound) and refused to compete against one another in reducing the spreads. But, for most currency pairs, bid–offer spread has come back to the pre-crisis level. Figure 10.1 Bid-offer spread of 3-month forward contracts in four currency pairs 2007–2010 0.20%
Composite (3-months Moving Average) EURUSD
0.18%
GBPUSD
0.16%
USDCHF
0.14%
USDJPY
0.12% 0.10% 0.08% 0.06% 0.04% 0.02% 0.00% 2007
2008
2009
2010
Source: Reuters Ecowin Pro
10.2.3 Other factors impacting currency liquidity 2000–2010 The following factors are important for an understanding of the wax and wane in foreign-exchange liquidity; many of these factors are linked to the proliferation of electronic dealing methods.
398
CO2_with_index_2nd_amends.indd 398
23/05/2011 14:09
Liquidity, Counterparty Risk and E Commerce
o High-frequency and algorithmic trading: It now accounts for a significant, and rising, portion of all foreign-exchange transactions. Many of the arbitrage opportunities are now less profitable and hence some the high-frequency traders have started providing liquidity to the market. Due to their presence, many market-making banks are much more cautious in terms of providing liquidity to their customers and they are making more effort screening their customers, mostly using the new technologies. o Greater acceptance of currency as a separate asset class: The currency-overlay industry was one major beneficiary of this trend over the decade. Trading volume from this industry increased significantly, largely due to the expansion of their pure-alpha business with much higher turnover than the passive and active overlay business. But in this space there are also many hedge funds, CTAs and traditional long-only funds. o “The Great Moderation”, dual currency deposits and leveraged carry positions: From 2004 up to the onset of the global financial crisis, foreign-exchange volatilities were hitting fresh historically low levels and heading lower. Many attributed this to the then benign global economic environment: high growth coupled everywhere with low inflation, so-called “Great Moderation”. Two types of trade were popular then, both among professional investors and retail investors. One is “dual-currency deposit”, where the investor puts money in deposit in one currency while selling a put option on another currency; hence, for enhanced yield on the deposit, the investor has the obligation at maturity to receive the weaker of the two currencies. Another type was the good old carry trade. Low volatility in the currency market encouraged many investors to increase leverage. Both trades contributed to the level of liquidity in the foreign-exchange market and, later, when investors took a collective and chaotic exit, they also helped to push the volatility to a very high level within a very short time frame. o Lehman Brothers and Bear Stearns: It was a shock to the foreignexchange market but the trading volume in spot had actually increased as a result. Many other markets froze and the currency market was still operating, if with higher frictions as the bid–offer spread went up by four- or fivefold during the height of the 399
CO2_with_index_2nd_amends.indd 399
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
crisis, some investors used currencies as proxies for other asset classes (King and Rime 2010). One lasting lesson from this episode was to focus investors’ and overlay managers’ attention on the credit risk management issues. o Reserve accumulation: The amount of foreign exchange reserves accumulated by central banks has been setting record highs quarter after quarter in recent years. What reserve managers may do with their massive pile of US dollars has a major psychological impact on the foreign-exchange market. This is one of the factors contributing to the fact that the change of liquidity situation is becoming a lot more abrupt. o Retail foreign-exchange traders: Though a long established battlefield for some retail investors, foreign-exchange market is luring more of them via the new generation of user-friendly online dealing tools, tight bid–offer spread and high-margin trading limits. The players offering retail foreign-exchange trading services are called “retail aggregators” and they include FXCM and FX Dealer Direct (King and Rime 2010). o Bank consolidation and specialisation: This point is discussed below. The consolidation in foreign-exchange markets is still ongoing at the time of writing. In 2001, around 16 top banks in large foreignexchange centres around the world accounted for 75% of the turnover in those centres respectively, according to King and Rime (2010). In the 2010 survey, around eight top banks would share between them 75% of the total business. Many small, regional and niche banks have outsourced their trading activities to large investment banks and concentrate instead on client deals. This is reflected in the growth of what is known as the “bank for bank” (or B4B) segment, which also serves the retail aggregators. At the same time, large banks have been spending money on upgrading the IT infrastructure and capture economies of scale. The fixed cost is very high, hence the high barrier to entry. But for those who have managed to build global infrastructure in terms of electronic trading and all auxiliary support systems, the reward has been handsome. Those large banks have huge deal flows and huge amounts of information on their customers, not unlike the Internet search company Google on its users. Some banks are said to use 400
CO2_with_index_2nd_amends.indd 400
23/05/2011 14:09
Liquidity, Counterparty Risk and E Commerce
algorithms to mine all the information and keep certain “scoring systems” on their customers. The ones good at making money are the bad ones for the banks, and their trading positions are routinely hedged out in the interbank market. Those prone to conduct lossmaking trades are actually the good ones for the banks, and their positions are kept on banks’ own books. Some customers also complained that banks have implemented algorithms to “learn” how they trade and adjust the pricing or their own positions accordingly. Hence, for many currency managers, execution management becomes a tougher job, despite the easy tools and ample liquidity. Some banks offer the algorithmic countersolution for the problems caused by algorithms: they offer algorithms to help currency-overlay managers manage their execution process. These algorithms may break large deals into smaller chunks; some chunks may even have opposite signs, and feed to the market in quick-fire fashion in order to achieve better execution prices. 10.3 COUNTERPARTY CREDIT RISK Historically, both the counterparty banks and the investors (and their currency-overlay managers) viewed credit risk embedded in short-term currency transactions as more of a theoretical risk factor than a practical one. Settlement risk was deemed more relevant and serious. Also, historically, only counterparty banks had some procedures in terms of classifying different types of investors, granting them different amounts of credit lines for trading, requiring some investors to post collaterals, and reviewing the outstanding credit exposures to a single investor after certain transactions had taken place. Investors and currency-overlay managers would just choose a handful of well-recognised names as counterparty banks and that was all in terms of counterparty credit risk management. The increasing level of adoption of prime-brokerage services and increasing use of sophisticated online tools in foreign-exchange trading made the situation more complicated for both counterparty banks and investors. For the banks acting as prime brokers, they have to monitor and manage a complex web of given-up transactions. They also need to manage counterparty credit risk from a range of different venues: traditional direct deals by telephone, on single bank platforms, or via many multibank platforms. For 401
CO2_with_index_2nd_amends.indd 401
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
investors using prime-brokerage services, their credit risk is now concentrated with one single prime broker. If for some reason the relationship ends, the investor faces significant risk of business disruption as well as potential losses on their trading portfolio and collaterals. Online tools may allow a currency-overlay manager to deal on behalf of several currency-overlay investors in one transaction and the counterparty bank will need to monitor the credit risk of a group of investors in a single transaction. A currency-overlay manager using a multibank platform may hit a trade button without knowing the exact counterparty bank behind the price quote (though the manager can specify beforehand which banks are on the dealing panel). All these make the issue of credit risk management more complicated. According to Pichardo-Allison (2009), the downfall of so many large financial institutions has made protecting against currency risk a risky endeavour. As a result, both investors and currencyoverlay managers have been adding or strengthening their internal policies dealing with counterparty credit risks. A perfect solution does not exist, but here are several practical methods investors and currency-overlay managers have been using to deal with the issue. o Monitoring counterparty credit quality more closely: This can be achieved by setting a minimum level of credit rating for counterparty banks. Some currency-overlay managers check more dynamic “credit default spreads” (CDSs) rather than just relying on more static credit rating by rating agencies. Some currencyoverlay managers conduct more thorough and frequent due-diligence exercises on the counterparty banks. o Capping maximum exposure to a single counterparty bank: If a prime-brokerage service is not used, a panel of counterparty banks is used and a cap of maximum exposure to a single bank can be put in place. o Implementing a second (or even third) prime broker, using a third-party account for collaterals. The many benefits of using a prime-brokerage service for currency overlay programmes need to be balanced against concentration of credit risk, especially if currency-overlay managers also leave all excess cash balances in deposit with the prime broker. If possible, all collaterals and cash 402
CO2_with_index_2nd_amends.indd 402
23/05/2011 14:09
Liquidity, Counterparty Risk and E Commerce
o
o
o
o
balances should be held in a third-party account, for example, with a custodian. There are cases for two-way collateral arrangement where counterparty banks need to post collateral to investors under certain conditions but those are exceptions rather than the market norm. A second prime-broker relationship should be in place so a switch can be implemented swiftly when needed. Putting in place a formal counterparty risk-management policy: This involves detailing approved and disapproved counterparty banks, minimum credit rating, action on downgrade, maximum allocation, collateral management, cash balance management and so on. Using futures rather than forwards: The volume of exchangetraded foreign-exchange futures has been grown at a much faster pace than the overall market turnover. While the main driver has been algorithmic/high-frequency trading, the concern of counterparty risk has driven some of the foreign-exchange deals to the exchanges. There are currently also several initiatives to centrally clear all OTC foreign-exchange transactions in the future. This area will be one of the most important developments in foreign-exchange market over the next few years. Using shorter maturity forwards: Everything else being equal, credit risk is smaller for forwards with shorter maturities. But this should be an emergency measure, as shorter maturity may result in higher cost, as discussed in Chapter 3, Section 3.3.5. Using the CLS service: As explained later in Section 10.4.3, CLS can eliminate settlement risk.
10.4 E-COMMERCE 10.4.1 General observations Currency trading is by far the most “electrified” capital market activity in both depth and width. Section 10.2.2 gave an overview of the extent of automation in the execution process, but in recent years the process of automation has spread to other aspects of currency trading, from pre-trade to at-trade to post-trade – in other words, to the whole process. At the pre-trade stage the delivery of research materials has gone electronic: there is now a whole range of Internet-based tools that allow the user to conduct ad hoc studies 403
CO2_with_index_2nd_amends.indd 403
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
with an increasing degree of flexibility and sophistication. At the at-trade stage, automatic liquidity and execution are now taken for granted, and the new standard is on whether the trades are electronically captured by the different systems used by clients. The post-trade stage is being revolutionised by automation and connectivity: STP, prime brokerage and CLS are technologies and methods that are bringing significant efficiency gains to end-users and are increasingly being adopted by different clients. 10.4.2 Straight-through processing Traditionally, foreign-exchange deals between a bank and its customers have been executed over the phone. Once the details are agreed, they are entered manually by both institutions into their respective systems. This process is error-prone, given that dealers are usually under time pressure. And errors can be costly in the fast-moving foreign-exchange markets. Once the deals are captured in the systems, they are processed further by middle- or back-office staff to add confirmation and settlement details. The process is largely automated now, but manual intervention is frequently required, which again can result in errors. Settlement errors can be expensive too because of the related interest charges and penalties. The emergence and increasing acceptance of online dealing platforms over the last five years is rapidly changing the way in which foreign-exchange transactions are executed. Many asset managers have signed on to proprietary systems provided by banks or multibank portals or both. Many of these systems enable them to automate the dealing process, reduce errors and achieve best execution. Clients can choose to integrate their own trading and back-office systems with these systems and achieve straight-through processing. The integration process is becoming standardised and the cost has been falling rapidly over recent years. All these factors have contributed to an increasing acceptance of online trading systems and straight-through processing. In the next two sub-sections we will look at examples of multi- and single-bank systems. 10.4 3 Continuous linked settlement CLS Bank was launched in 2002 by the world’s largest banks and financial institutions to create the only global settlement system 404
CO2_with_index_2nd_amends.indd 404
23/05/2011 14:09
Liquidity, Counterparty Risk and E Commerce
for managing settlement risk. CLS Bank allows members and their clients to settle payment instructions related to foreign-exchange deals in real time, eliminating the time-zone differences that can cause payment delays and give rise to settlement risk. Settlement risk is also known Herstatt risk, referring to the failure of Herstatt Bank in 1974, when it had received its foreign-currency payments but went bankrupt before making its US dollar payment, resulting in losses to its counterparties. Settlement risk refers to the risk of loss when one party makes the payment in one currency as part of a foreign-exchange transaction but does not receive the payment in another currency. Historically, currency payments are settled during different time periods in 24 hours of a single day, one currency typically settled during the trading hours of the country that issues this currency. Time-zone difference can therefore cause settlement risk. CLS settles the trades in a continuous process, matching one payment in one currency against the other payment in another currency simultaneously, therefore eliminating the settlement risk. The existence of CLS was credited for avoiding major market interruptions following the demise of Lehman Brothers, which was a major counterparty bank in foreign-exchange market (King and Rime, 2010). Most developed currencies and a few emerging-market currencies are included in the CLS service (17 currencies at the time of writing) and up-to-date information can be found on its website http://www.cls-group.com. Currency-overlay managers can participate in CLS if the custodian and the counterparty banks are CLS members and the currencies are among the CLS included currencies. The take-up rate for CLS among currency-overlay managers is still low but is increasing. In addition to reducing settlement risk, CLS can also improve straight-through processing and thus improve operational efficiency. It charges a small fee to both parties of the transaction.3 10.4.4 Multibank platforms There are now several main multibank platforms: for example, Currenex, FXConnect, FXall, Hotspot and FXCM. New entrants are still appearing in this market, usually focusing on some niche areas. These platforms allow different market makers to quote prices in 405
CO2_with_index_2nd_amends.indd 405
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
competition with one another.4 The main advantage of a multibank platform lies in competitive pricing and it is ideal for currencyoverlay managers with a fiduciary responsibility to obtain the best execution. However, all platforms have expanded (and continue to expand) their services to include pre-trade and post-trade activities. Here we use FXall as an example to explain what a multibank platform offers. The offerings of different platforms are usually only marginally different. FXall reached US$125 billion daily trading volume in November 2010 and was voted the number-one multibank platform by Euromoney and FX Week for nine and eight consecutive years respectively. Like other online multibank portals, FXall’s value proposition as a financial intermediary can be categorised as follows. o Trading: Trading is the core competency and FXall has a variety of execution mechanisms suitable for different trading requirements. For example, if you choose to use QuickTrade method (request for quote), FXall requests prices from participating banks according to the client’s specifications. The prices are displayed and best prices highlighted. The client needs only to “click and deal” within a specified time. The price-making process is highly automated and done within a split second. o Operation: Modules allow pre- and post-trade allocation and amendment. o Settlement: The deal-matching, confirmation and settlement process is also automated between the liquidity-providing banks and FXall through STP. The client can opt to integrate its own systems fully with FXall to realise the full potential of online dealing with STP. The integration service is provided by a host of system providers partnered with FXall. o Information: Aside from identifying the best-quoting banks in real time, FXall can also measure the performance of chosen banks. It also offers an “Execution Quality Analysis” service, analysing investors’ trading over time according to several metrics in order to help clients assess the effectiveness of their current execution strategy. The trading platform’s execution methods are designed to im406
CO2_with_index_2nd_amends.indd 406
23/05/2011 14:09
Liquidity, Counterparty Risk and E Commerce
prove efficiency. Investors can choose different execution mechanisms including collaborative dealing, benchmark fixing, requests for quotes, continuous streaming prices (bank streams), and order book – screenshots of some of these execution modules are shown in Figure 10.2 and further details can be found on the company’s website at http://www.fxall.com. A greater degree of efficiency than the traditional dealing channels can be achieved by reducing the time to price small trades through automation; by standardisation of confirmation and settlement; and by centralisation of training and support. In contrast to the online dealing tools provided by banks, FXall’s services and products, which are provided by a third party, help to avoid potential conflicts of interest by providing fullcompliance workflow controls and audit trails. Figure 10.2 Screenshots of the FXall trading platform
Source: FX Alliance LLC, available at http://www.fxall.com
Other benefits, no less important, lie in the automation of other steps of a deal’s life cycle. For example, FXall’s settlement centre has the following capabilities:
407
CO2_with_index_2nd_amends.indd 407
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
o automated capture of trade details, creating an audit trail and providing data to benchmark execution quality; o immediate and constant matching of transaction details, netting, custodial notification, prime-broker give-ups, and multi-account allocations; o user-based permissioning to enforce segregation of duties; o real-time credit monitoring to help you manage counterpart risk; and o CLS capabilities – decision/routing logic to identify CLS, clearing eligible trades. 10.4.5 Single-bank platforms Nearly all large foreign-exchange banks offer eFX solutions of one form or another. Apart from the fact that just one bank provides the dealing prices in these systems, their functionality is fairly similar to that of multibank portals. Because the number of such systems is much larger than the number of multibank portals, they do vary in the range and depth of services provided. Here we offer a general overview of such facilities using UBS’s online system as an example. UBS’s e-tool suite is designed to address the whole trading process from pre-trade to post-trade. It includes: o pre-trade (decision support): o FX Web (which includes online foreign exchange research and analytics); o Interchange Chat (instant messaging system); o at-trade (liquidity): o FX Trader Plus (spot, forward, swap, NDFs, precious metals, vanilla options; streaming dealable prices; one-click trading; order management); o post-trade: o FX KeyLink (confirmation, settlement and payment module); o FX2B (system-to-system integration for banks, corporations and intermediaries where UBS provides liquidity to clients’ front-end systems. Also referred to as “white label” foreign exchange liquidity). 408
CO2_with_index_2nd_amends.indd 408
23/05/2011 14:09
Liquidity, Counterparty Risk and E Commerce
The website is designed to reduce information overload. A variety of shortcuts, tags and filters (which online technology is most suited to delivering) are used to assist users to locate relevant information with speed and accuracy. The commentaries can be distributed to clients via email, Bloomberg, chat, web, or third-party channels (such as FXall). Its Analytics section allows users to access a range of data-mining tools for analysing historical and real-time data – in particular UBS’s proprietary implied volatility data. The website also offers a black-box-driven technical trading model complete with data on positions and historical performance. Figure 10.3 Screenshots of UBS’s online eFX tool suites
Source: UBS
The execution modules provide firm, dealable prices for a range of products on a click-and-deal basis. “Firm” means that prices cannot be changed or revoked over a specific time window when they are sent to the client for acceptance or rejection. According to UBS, most deals going through its dealing modules do not require dealer intervention. Figure 10.3 has screenshots of some different modules. The post-trade module, KeyLink, can handle a variety of func409
CO2_with_index_2nd_amends.indd 409
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
tions, including trading, confirmation, cash management and security management, across different asset classes. Because it is SWIFT-based, clients can use it to access confirmation and payment information from any SWIFT-based bank. 10.4.6 eFX and currency management in practice For a currency-overlay manager the main and most direct benefit of eFX is the reduction of error and other types of operational risk. Other direct benefits of eFX include better price discovery, the availability of audit trails, lower overheads and better management of credit relationships. Because the whole dealing process is done electronically, audit and reporting capabilities are also greatly increased. During the early part of 2000s, many currency managers expressed reservations on the likely impact of eFX on their business. While most acknowledged the potential for error reduction through STP, there was concern about the cost of integration. They were even less enthusiastic about other benefits the online foreign-exchange paradigm might bring. However, the situation has changed significantly, with sell-side institutions – including banks and multibank portals – making rapid progress on the services they provide. Technological advances, coupled with alliances between liquidity providers and system providers, are driving down the cost of integration while increasing its scope. The previous discussion on eFX focuses on the online foreignexchange transaction and the steps immediately before and after the point of transaction. More recently, eFX has been penetrating many other aspects of currency management, for instance: o exposure identification and aggregation, especially the information flow between the custodians and the currency-overlay managers; o allocation of trades to clients’ accounts and averaging of prices of multiple transactions (these are very important functions for currency-overlay managers); o portfolio construction according to individual constraints; o risk control; o cash management; and o performance measurement and real-time reporting.
410
CO2_with_index_2nd_amends.indd 410
23/05/2011 14:09
Liquidity, Counterparty Risk and E Commerce
Many currency-overlay managers still use legacy systems or systems primarily designed for managing other asset classes; newer systems typically make good use of the latest features of technology and automate the following aspects in a currency-overlay process: o collect exposure data from underlying portfolios within a short period; o compute passive currency-hedging requirements using inbuilt hedge ratios relevant for each account and taking into account the specific constraints; o implement active currency decisions across different accounts under their individual risk limits and constraints; o aggregate net currency positions and generate the new trades required; o check counterparty credit limits; o route the new trades through single- or multibank platforms on the dealer’s instruction; o report back prices and, if they are accepted by dealers, capture the trades in internal risk-control systems; o confirm and settle all trades and/or, if required, send details to custodians; and o generate reports for different purposes. Good systems that perform all the above tricks seamlessly are still exceptions rather than the norm. Most currency managers still carry different IT baggage due to corporate restructuring and/or a lack of investment/initiatives. This aspect of currency-overlay management will see the most significant changes in the next few years. 10.4.7 Custodian banks and eFX The market for custodians has consolidated at some pace in recent years, with many banks exiting the business and those remaining able to exploit economies of scale. Many of the custodians offer online services to investment managers that cover both execution and risk management solutions. The change is partly driven by the demand from investment managers, who increasingly view foreign exchange as an asset class for returns and want to consolidate foreign-exchange trading and reduce transaction costs. 411
CO2_with_index_2nd_amends.indd 411
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
Some custodians are offering what is known as 360-degree STP. This automates several tasks in the overlay process, including exposure identification and consolidation, price discovery (execution), confirmation and settlement, and position reporting and risk analysis. The STP-enabled overlay process is significantly simplified. The custodian bank acts as a hub for most of the STP links, automatically collating information at pre-, at- and post-trading stages. The information can be processed promptly and accurately and provided to both fund managers and overlay managers to act on.
1. Market makers “call out” when they try to obtain prices, usually simultaneously, from other market makers. They then choose to deal one or several of the obtained prices – in other words, they “hit” on these quotes. A call-out offers one of the most entertaining glimpses of life on a foreign-exchange trading floor, where several traders punch furiously at their dealing terminals and shout prices or instructions at one another. Its occurrence is now increasingly less common. 2. Lyons’s website contains more up-to-date information: see http://faculty.haas.berkeley. edu/lyons/index.htm. 3. See DuCharme (2008) for further details. 4. See Gallardo and Heath (2009) for a detailed survey of execution methods.
412
CO2_with_index_2nd_amends.indd 412
23/05/2011 14:09
11
Conclusions
11.1 SUMMARY OF THE MAIN ARGUMENTS We set out to survey and explain the concepts and practices of currency overlay in a neutral, practical and informative manner, but we ended up with a book that is probably much broader in its scope. The expansion was, in our opinion, necessary to achieve the original objective. The variability in investment returns caused by currency movements is an unavoidable and direct consequence of investing globally. Additional variability in returns can also be introduced indirectly as a consequence of the globalisation of both the goods and services markets and the financial markets around the world. Some investors treat such variability as a risk factor while others do not. Those who do not view currency as a risk factor often liken currency returns to a random walk whose long-term impact on investment returns will wash out. Among those who do treat currency variability as a risk factor, some also subscribe to the random-walk hypothesis. For them, predicting currency returns is all but futile and they usually opt for some form of passive hedging to remove part or all of the return variability caused by currencies. Some, on the other hand, believe that currency markets are inherently inefficient and that currency movements can be predicted to a certain extent. They therefore view currency variability as a profit opportunity in that positive impact can be retained while negative impact is eliminated. This divergence in investors’ views on the existence of currency 413
CO2_with_index_2nd_amends.indd 413
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
risk provides a rich and complex background against which the question of currency overlay is discussed. For many investors, or for those involved in currency risk management, these views have typically formed over a long period of time and have, so to speak, ossified. Discussion of currency overlay is often sidelined by the much broader question of how to deal with currency risk in an investment portfolio. We therefore started our exploration from the standpoint of this second question. We argued that, no matter what view an investor might have on whether currency movements represent a random walk (or any of the questions covered in Chapter 1, Section 1.5), the currency component of a portfolio should be separated from other risk factors – for example, risks in bonds and risks in equities. Our view on risk separation provided us with a framework for surveying the theoretical as well as the practical aspects of currency overlay, which, according to our definition, represents nothing more than a process of separating the variability of returns caused by currency moves. The reason for our belief in currency separation – and our favourable disposition to currency overlay – lies not so much in our view on the predictability of currency movements, a topic on which we spent some time (in Chapter 4 and parts of Chapters 7 and 8) reviewing the existing evidence; nor in the body of empirical work that generally favours currency management (Chapter 1, Section 1.6.4); nor in the track record of the specialist currency-overlay managers or their well-organised market presence (Chapter 2). Our belief in currency separation stems from a simple premise: any investment should be managed by a manager with a declared core competency in managing that particular type of risk factor or factors (Chapter 2, Section 2.4.4). If currency fluctuation, be it a risk factor or not, does not fall under a fixed-income or equity manager’s core competency, then currency impact should be excluded from that manager’s remit, ie, the manager’s performance should be measured on a currency-neutral basis. Currency-neutral here means that the manager’s decision-making or performance-measurement processes should not be influenced by the fluctuations in currencies. Even in cases where the underlying managers do declare additional core competency in currencies, they should still be measured on a currency-neutral basis for the underlying and their currency decisions should be evaluated separately. 414
CO2_with_index_2nd_amends.indd 414
23/05/2011 14:09
conclusions
Although we promote currency-neutral asset allocation and a performance-measurement framework (Chapter 6), we do not encourage an investment manager to actually carry out the full currencyhedging transaction, nor do we discourage them from doing so. In fact, we have no fixed view on particular hedge ratios (Chapter 5), active or passive hedging styles (Chapter 3, Section 3.2), or specific trading models (Chapter 4). These are not necessarily contradictory statements. The fully currency-neutral framework will ensure that investment decisions are separated, at least conceptually. What does an investor do with the separated and centralised currency part? We think the decision should be made in more or less the same manner in which the investor makes other choices: bonds versus equity, domestic versus overseas, growth versus value, index versus active, etc, typically via an optimisation process riddled with practical fudges to achieve a desirable risk/return objective. We have presented factual information on the risks and returns of currencies, as well as on methods for quantifying these risks and returns (mainly in Chapter 1, Section 1.3, but also in Chapters 4, 6, 8 and 9), which we hope can be used as input variables in the investor’s decision-making process. We also discussed, at considerable length, some specific issues related to currency and optimisation techniques, including the benchmark hedge ratio (Chapter 5). If the investor’s currency decisions entail more than just passive hedging, they can entrust the currency decisions to managers who have a core competency in currency management. However, we need to emphasise that the decision on the benchmark hedge ratio is potentially a much more important (Chapter 7) and difficult one to outsource. We propose a practical solution to this conundrum in the next section. Of course, the above picture – where an investor performs optimisation, takes currency decisions and selects managers according to their core competency – is a gross oversimplification of the real-life situation where the relationship between investor and investment managers is much more complex, as are the decisionmaking processes. This is particularly the case in large institutions. We feel that it is precisely in this type of complex situation that the currency-overlay process can bring substantial, if less quantifiable, benefits. In order to run an overlay efficiently, an investor or invest415
CO2_with_index_2nd_amends.indd 415
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
ment manager will likely need to know, with a certain degree of accuracy and timeliness, their investment positions in nearly all asset classes so that they can figure out the currency exposures on which an overlay programme will run. The process of collecting, distilling and sharing this data, together with clarified risk responsibility, will undoubtedly help to improve the overall efficiency of investment institutions through the clarification of risk responsibility. In the previous edition, we said that we were of the opinion that the (external) specialist overlay industry would continue to grow, and it has been doing so. We think that future growth in this industry will still be healthy, if not stellar. We continue to believe that the main growth area will be organic, in that institutional managers will set up their own, internal overlay processes in some shape or form, not only to seek returns and/or reduce risks but also exploit the informational benefit of currency overlay. The other area where the business is set to grow is from the institutional investors based in emerging markets. The cost of setting up a currency-overlay programme will continue to fall, largely thanks to the proliferation of online dealing and risk-management tools (see Chapter 10). Many institutional fund managers are taking steps to realise the full potential of e-commerce in a currency-overlay process and this will continue. And how to deal with emerging-market currencies exposures is a topic that the industry will have to deal with in the next decade (see Chapter 9). 11.2 A PRACTICAL FRAMEWORK OF SETTING BENCHMARK HEDGE RATIOS We have discussed in several sections of this book that setting a benchmark hedge ratio would probably make the largest impact in terms of the risk-and-return profile of a portfolio over both the short and long terms. In addition, it is very difficult to outsource this decision to external fund managers due to practical constraints such as the performance review cycle, the potential reaction of investors towards persistent negative returns and the fund manager’s career horizon. There are many ways of setting the benchmark hedge ratio for a particular portfolio and/or investor. Figure 11.1 summarises an optimisation approach with practical fudges: you use a historical variance–covariance matrix between currencies and 416
CO2_with_index_2nd_amends.indd 416
23/05/2011 14:09
conclusions
underlying assets in a standard portfolio-optimisation procedure, aiming to minimise portfolio risk (or, alternatively, to maximise the risk-adjusted portfolio return). The optimisation is usually constrained by practical considerations such as the beneficial aspect of currency exposures in liability hedging, or the allocation to foreign assets relative to the total portfolio, or the cashflow management concerns, etc. This optimisation-based approach offers some practical guidance in terms of selecting a benchmark hedge ratio but is probably of limited value, mostly due to the instability of the correlation structures. Hence, in real life, many investors will go through the optimisation but pick a benchmark hedge ratio bearing no relationship to the results from optimisation. Figure 11.1 Using optimisation together with practical constraints to set benchmark hedge ratio Input •
Base currency and exposure currencies
•
Currency volatities and correlations
•
Underlying assets types, volatilities and correlations
•
Time window for historical data
% of foreign asset allocation
Emerging markets currencies
Optimisation procedure to determine hedge ration
Result •
Volatility: risk reduction up to some extent
•
Return: performance enhancement quite debatable
Different hedge ratios for different asset classes or overall hedge ratio
Asset/liabilities matching
Cashflows
It is typical to see that many investors set and/or change their benchmark hedge ratios reacting to the large and/or persistent moves in the base currency relative to exposure currencies, by which time they would have suffered either a heavy accounting loss on the underlying exposures if they did not hedge or heavy cash outflows on the hedging portfolios if they did hedge. This is depicted in Figure 11.2. The problem of this approach is obvious: 417
CO2_with_index_2nd_amends.indd 417
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
the investors are always chasing the market, are prone to pay twice (see the empirical evidence on this point in Brandes Institute 2007), and the hedging decisions are made with little consistency. Figure 11.2 Reactive and proactive approaches to setting a benchmark hedge ratio Proactive
Reactive
• Action when
• Action when
* As part of review cycle * Investors, in consultation with FX
* Having suffered large losses on previous hedging decisions
specialists and consultants
* Investor acts alone
• Cons * Always chasing market * Often pay twice * Lack of consistency • Pros * Potentially quicker
VS.
• Cons * Decision that takes more time * Complex framework, does not always lead to better results • Pros * Clarify risk responsibility * Consistency
Instead, we believe that an investor should choose a proactive approach in setting the benchmark hedge ratio and this approach should: o be part of the review cycle of the whole investment approach and asset allocation; o have a long horizon, at least between three and five years, if not longer; o have very infrequent adjustments to the chosen benchmark hedge ratio, unless major moves have occurred in the base currency and its departure from fair-value is significant; and o have a baseline benchmark hedge ratio around the minimum regret hedge ratio, 50%, adjusted up and down according to: o relative allocation to foreign-currency-denominated assets – if low, then benchmark hedge ratio should be lower; if high, higher; o fair value of the base currency against the major exposure currencies – if there is major undervaluation, then benchmark hedge ratio should be higher; if there is major overvaluation, lower; under- and overvaluation could be created after a major 418
CO2_with_index_2nd_amends.indd 418
23/05/2011 14:09
conclusions
move in base currency; implicitly, the adjustment is contrarian; o possibly, the differential in long-term productivity growth – if productivity growth is persistently higher for domestic economy versus foreign economies, then the benchmark hedge ratio should be higher; if lower, lower; o possibly, the forward bias if it is persistent for investor’s base currency – if persistent positive carry, then benchmark hedge ratio should be higher; if persistent negative carry, lower. o possibly, the overall volatility of investor’s base currency against exposure currencies – if relatively high volatility, then benchmark hedge ratio should be higher; if low, lower. The framework can be clearly stated and communicated. Any short-term noises and timing issues should be exploited by using currency pure alpha or active overlay programmes in addition to the passive currency overlay, potentially with a mix of different styles and managers. This is shown in Figure 11.3. Figure 11.3 Illustration of how long-term (strategic) and short-term (tactic) currency decisions can be made with reference to each other
• Market inefficiencies • Opportunities in trend-following, carry and volatilities
• Specialist approach • Either pure alpha or an active currency overlay strategies
• Risk reduction • Currency misalignment • Long term productivity growth differential
• Passive overlay strategy • Infrequent change to the passive hedge ratio • Internal or external
• Short-term: Return enhancing • Longer-term: Risk reduction, and capturing longer trends if there are any
• The different kinds of FX risk are assumed and managed by people who are best positioned to manage them
Overall
Short term
Ways to exploit them
Long term
Opportunities
This way, the risk-reduction objective is achieved through a passive overlay, with infrequent changes to the benchmark hedge ratio in order to take advantage of occasional major misalignment in currencies. This latter point was well illustrated in “Why is it so difficult to beat the random walk forecast of exchange rates?” by Kilian and 419
CO2_with_index_2nd_amends.indd 419
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
Taylor (2001), and in “Currency Hedging Programs: A Long-Term Perspective” by Brandes Institute (2007). There is substantial risk in taking such decisions as exchange-rate tends to overshoot and, in our view, the investor, rather than the currency overlay managers, is best placed to shoulder this risk for long-term returns and risk reduction. Investor should avoid adjusting benchmark hedge ratio over a medium horizon, eg, one to three years, as plenty of evidence shows that it is mostly futile trying to outsmart the market over this horizon. Over shorter horizons, specialist currency managers are shown to be able to add some alpha. Useful, but only marginally. 11.3 A CURRENCY OVERLAY CHECKLIST The book being a practical guide, it probably makes sense to conclude it with a checklist for a hypothetical investment manager looking to start a currency overlay. Here it goes. 1. Forming a strategic view on currency risk
Does currency risk exist for your portfolio ? No
Clearly state/communicate the conclusion in the investment policy Yes
o Does it exist, and does it depend on the special circumstances of the portfolio/investor in question? o If so, to what extent? If not, why not? o What is the relevance of the questions asked in the hedging debate, particularly with regard to the expected returns from currency trading, to this view? o Is the risk different for bonds and equities? o Should emerging-market currencies be treated differently from developed-market currencies?
Should the risk be passively hedged?
420 No
Clearly state/communicate the conclusion in the CO2_with_index_2nd_amends.indd 420 investment policy, passively monitor
23/05/2011 14:09
Yes
conclusions
2. Deciding on passive hedging Should the risk be passively hedged? No
Clearly state/communicate the conclusion in the investment policy, passively monitor currency risk.
o What are the costs and ben efits of passive hedging? o Is the risk reduction prop erly quantified? o Is the fund’s currency ex posure expected to increase over time?
Yes
3. Selecting a passive hedge ratio
Which hedge ratio?
Can currency moves be predicted? No
Consider external overlay managers or stay with internal passive overlay. Consider risk reduction motivated hedging techniques.
Yes, go active
o What are the alternatives? o What are the pros and cons of each alternative in theory and in practice? o Should we take forward bias and/or PPP into consideration? o What are the problems of using optimisation techniques to determine a hedge ratio? Should the ratio be differo ent for bonds and equities? Should the execution of paso sive hedges be centralised? 4. Forming a strategic view on currency returns o Do they follow a random walk? o What is the evidence? o Do emerging-market cur rencies behave differently? o What magnitudes of return and risk are expected?
421 Which trading style(s)?
CO2_with_index_2nd_amends.indd 421
23/05/2011 14:09
No
Consider external overlay managers or stay with internal Yes, go active passive overlay. Consider risk CURRENCY OVERLAY – SECOND EDITION reduction motivated hedging techniques.
5. Selecting a trading style
Which trading style(s)?
Select trading style(s), back-testing performance and portfolio fit if needed
6. Reviewing managers’ core competency
Using internal or external managers? External
Selecting external managers by track record and other factors. Deciding on risk control parameters
o What are the relative strengths and weakness of different foreign-exchange determination models? o Are the results dependent on investment horizons? o Does style diversification work in currencies? o Should we also look at volatility-determination models?
Internal
o Do the investors let the underlying managers manage currency? o Should we employ currency specialists? o Who are the external managers and how do they differ from one another? o Should part of the exposure be given to an external manager in order to benchmark the internal manager’s performance? 7. More choices
Risk control and other practicalities
o Which hedging instruments? o Which tenor? o Which risk limits? o Which adjustment style? o Which rebalancing frequency? o How to manage cash balances? o Which counterparties?
422
CO2_with_index_2nd_amends.indd 422
23/05/2011 14:09
conclusions
o Should we use a prime bro ker? Electronic dealing? Single-bank or multibank platforms? 8. Evaluating before and after
Performance measurement
o What are the main issues in evaluating currency per formance? o What are the strengths and weaknesses of the different approaches? o How to construct a holding based currency benchmark? o How to use a performance based benchmark? o What are the contributions, in both return and risk terms, of passive hedging and ac tive hedging?
Clearer and cleaner investment policy; Competency-based approach to investment and risk management decisions; Better understanding of the portfolio risk as a whole and in its constituents; Centralisation of currency trading and other administration functions, with improved efficiency; Potential benefit of risk reduction so that the risk budget can be spent more productively; Potential source of alpha, uncorrelated with underlying returns
423
CO2_with_index_2nd_amends.indd 423
23/05/2011 14:09
CO2_with_index_2nd_amends.indd 424
23/05/2011 14:09
List of currencies
List of currencies
Currency names and codes vary between sources. Here in Table A.1 we list names and codes that are commonly used by foreign exchange market participants. Table A.1 contains a fuller list of currencies, threeletter codes, and locations that use the respective currencies. A full list of ISO currency names and codes (ISO 4217) can be found at http:// en.wikipedia.org/wiki/ISO_4217#cite_note-ReferenceA-4.
Table A.1 Currencies for which cash/derivatives products are commonly traded Currency
Code
Argentine peso
ARS
Australian dollar
AUD
Brazilian real
BRL
Bulgarian lev
BGN
Canadian dollar
CAD
Chilean peso
CLP
Chinese yuan*
CNY
Colombian peso
COP
Czech koruna
CZK
Danish krone
DKK
Egyptian pound
EGP
Euro
EUR
Hong Kong dollar
HKD
Hungarian forint
HUF
Icelandic króna
ISK
425
CO2_with_index_2nd_amends.indd 425
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
Table A.1 (continued) Indian rupee
INR
Indonesian rupiah
IDR
Israeli new sheqel
ILS
Japanese yen
JPY
Latvian lats
LVL
Lithuanian litas
LTL
Malaysian ringgit
MYR
Mexican peso
MXN
New Taiwan dollar
TWD
New Zealand dollar
NZD
Norwegian krone
NOK
Peruvian nuevo sol
PEN
Philippine peso
PHP
Polish złoty
PLN
Pound sterling
GBP
Romanian new leu
RON
Russian rouble
RUB
Saudi riyal
SAR
Singapore dollar
SGD
South African rand
ZAR
South Korean won
KRW
Special Drawing Rights
XDR
Swedish krona/kronor
SEK
Swiss franc
CHF
Thai baht
THB
Turkish lira
TRY
United States dollar
USD
Venezuelan bolívar fuerte
VEF
* The new offshore Chinese yuan (renminbi) is known as CNH since 2010.
426
CO2_with_index_2nd_amends.indd 426
23/05/2011 14:09
LIST OF CURRENCIES
Table A.2 Full list of currencies, including precious metals but excluding currencies of euro member states Code
Currency
Locations using the currency
AED
United Arab Emirates dirham
United Arab Emirates
AFN
Afghan afghani
Afghanistan
ALL
Albanian lek
Albania
AMD
Armenian dram
Armenia
ANG
Netherlands Antillean guilder
Curaçao, Sint Maarten
AOA
Angolan kwanza
Angola
ARS
Argentine peso
Argentina
AUD
Australian dollar
Australia
AWG
Aruban florin
Aruba
AZN
Azerbaijani manat
Azerbaijan
BAM
Bosnia and Herzegovina convertible mark
Bosnia and Herzegovina
BBD
Barbados dollar
Barbados
BDT
Bangladeshi taka
Bangladesh
BGN
Bulgarian lev
Bulgaria
BHD
Bahraini dinar
Bahrain
BIF
Burundian franc
Burundi
BMD
Bermudian dollar
Bermuda
BND
Brunei dollar
Brunei, Singapore
BOB
Boliviano
Bolivia
BRL
Brazilian real
Brazil
BSD
Bahamian dollar
Bahamas
BTN
Bhutanese ngultrum
Bhutan
BWP
Botswana pula
Botswana
BYR
Belarusian ruble
Belarus
BZD
Belize dollar
Belize
CAD
Canadian dollar
Canada
427
CO2_with_index_2nd_amends.indd 427
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
Table A.2 (continued) CDF
Congolese franc
Democratic Republic of Congo
CHF
Swiss franc
Switzerland, Liechtenstein
CLP
Chilean peso
Chile
CNY
Chinese yuan
China (People’s Republic)
COP
Colombian peso
Colombia
COU
Unidad de Valor Real
Colombia
CRC
Costa Rican colon
Costa Rica
CUC
Cuban convertible peso
Cuba
CUP
Cuban peso
Cuba
CVE
Cape Verde escudo
Cape Verde
CZK
Czech koruna
Czech Republic
DJF
Djiboutian franc
Djibouti
DKK
Danish krone
Denmark, Faroe Islands, Greenland
DOP
Dominican peso
Dominican Republic
DZD
Algerian dinar
Algeria
EGP
Egyptian pound
Egypt
ERN
Eritrean nakfa
Eritrea
ETB
Ethiopian birr
Ethiopia
EUR
Euro
17 European Union countries (Austria, Belgium, Cyprus, Estonia, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Malta, the Netherlands, Portugal, Slovakia, Slovenia, and Spain), Andorra, Kosovo, Monaco, Montenegro, San Marino, Vatican City
FJD
Fiji dollar
Fiji
FKP
Falkland Islands pound
Falkland Islands
GBP
Pound sterling
United Kingdom
GEL
Georgian lari
Georgia
GHS
Ghanaian cedi
Ghana
GIP
Gibraltar pound
Gibraltar
428
CO2_with_index_2nd_amends.indd 428
23/05/2011 14:09
LIST OF CURRENCIES
Table A.2 (continued) GMD
Gambian dalasi
Gambia
GNF
Guinean franc
Guinea
GTQ
Guatemalan quetzal
Guatemala
GYD
Guyanese dollar
Guyana
HKD
Hong Kong dollar
Hong Kong Special Administrative Region, Macau Special Administrative Region
HNL
Honduran lempira
Honduras
HRK
Croatian kuna
Croatia
HTG
Haitian gourde
Haiti
HUF
Hungarian forint
Hungary
IDR
Indonesian rupiah
Indonesia
ILS
Israeli new sheqel
Israel
INR
Indian rupee
Bhutan, India
IQD
Iraqi dinar
Iraq
IRR
Iranian rial
Iran
ISK
Icelandic króna
Iceland
JMD
Jamaican dollar
Jamaica
JOD
Jordanian dinar
Jordan
JPY
Japanese yen
Japan
KES
Kenyan shilling
Kenya
KGS
Kyrgyzstani som
Kyrgyzstan
KHR
Cambodian riel
Cambodia
KMF
Comoro franc
Comoros
KPW
North Korean won
North Korea
KRW
South Korean won
South Korea
KWD
Kuwaiti dinar
Kuwait
KYD
Cayman Islands dollar
Cayman Islands
KZT
Kazakhstani tenge
Kazakhstan
LAK
Lao kip
Laos
429
CO2_with_index_2nd_amends.indd 429
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
Table A.2 (continued) LBP
Lebanese pound
Lebanon
LKR
Sri Lanka rupee
Sri Lanka
LRD
Liberian dollar
Liberia
LSL
Lesotho loti
Lesotho
LTL
Lithuanian litas
Lithuania
LVL
Latvian lats
Latvia
LYD
Libyan dinar
Libya
MAD
Moroccan dirham
Morocco
MDL
Moldovan leu
Moldova (except Transnistria)
MGA
Malagasy ariary
Madagascar
MKD
Macedonian denar
Republic of Macedonia
MMK
Myanma kyat
Myanmar
MNT
Mongolian tugrik
Mongolia
MOP
Macanese pataca
Macau Special Administrative Region
MRO
Mauritanian ouguiya
Mauritania
MUR
Mauritian rupee
Mauritius
MVR
Maldivian rufiyaa
Maldives
MWK
Malawian kwacha
Malawi
MXN
Mexican peso
Mexico
MYR
Malaysian ringgit
Malaysia
MZN
Mozambican metical
Mozambique
NAD
Namibian dollar
Namibia
NGN
Nigerian naira
Nigeria
NIO
Cordoba oro
Nicaragua
NOK
Norwegian krone
Norway
NPR
Nepalese rupee
Nepal
NZD
New Zealand dollar
New Zealand
OMR
Omani rial
Oman
PAB
Panamanian balboa
Panama
430
CO2_with_index_2nd_amends.indd 430
23/05/2011 14:09
LIST OF CURRENCIES
Table A.2 (continued) PEN
Peruvian nuevo sol
Peru
PGK
Papua New Guinean kina
Papua New Guinea
PHP
Philippine peso
Philippines
PKR
Pakistani rupee
Pakistan
PLN
Polish złoty
Poland
PYG
Paraguayan guaraní
Paraguay
QAR
Qatari rial
Qatar
RON
Romanian new leu
Romania
RSD
Serbian dinar
Serbia
RUB
Russian rouble
Russia, Abkhazia, South Ossetia
RWF
Rwandan franc
Rwanda
SAR
Saudi riyal
Saudi Arabia
SBD
Solomon Islands dollar
Solomon Islands
SCR
Seychelles rupee
Seychelles
SDG
Sudanese pound
Sudan
SEK
Swedish krona/kronor
Sweden
SGD
Singapore dollar
Singapore, Brunei
SHP
Saint Helena pound
Saint Helena
SLL
Sierra Leonean leone
Sierra Leone
SOS
Somali shilling
Somalia
SRD
Surinamese dollar
Suriname
STD
São Tomé and Príncipe dobra
São Tomé and Príncipe
SYP
Syrian pound
Syria
SZL
Lilangeni
Swaziland
THB
Thai baht
Thailand
TJS
Tajikistani somoni
Tajikistan
TMT
Turkmenistani manat
Turkmenistan
TND
Tunisian dinar
Tunisia
TOP
Tongan pa‘anga
Tonga
431
CO2_with_index_2nd_amends.indd 431
23/05/2011 14:09
CURRENCY OVERLAY – SECOND EDITION
Table A.2 (continued) TRY
Turkish lira
Turkey, Northern Cyprus
TTD
Trinidad and Tobago dollar
Trinidad and Tobago
TWD
New Taiwan dollar
Taiwan
TZS
Tanzanian shilling
Tanzania
UAH
Ukrainian hryvnia
Ukraine
UGX
Ugandan shilling
Uganda
USD
United States dollar
United States, Ecuador, El Salvador, Guam, Haiti, Palau, Panama, Puerto Rico, Timor-Leste, Virgin Islands, Bermuda (as well as Bermudian Dollar)
UYU
Uruguayan peso
Uruguay
UZS
Uzbekistan som
Uzbekistan
VEF
Venezuelan bolívar fuerte
Venezuela
VND
Vietnamese
VUV
Vanuatu vatu
Vanuatu
WST
Samoan tala
Samoa
XAF
CFA franc BEAC
Cameroon, Central African Republic, Republic of the Congo, Chad, Equatorial Guinea, Gabon
XAG
Silver (one troy ounce)
XAU
Gold (one troy ounce)
XCD
East Caribbean dollar
Anguilla, Antigua and Barbuda, Dominica, Grenada, Montserrat, Saint Kitts and Nevis, Saint Lucia, Saint Vincent and the Grenadines
XDR
Special Drawing Rights
International Monetary Fund
XOF
CFA Franc BCEAO
Benin, Burkina Faso, Côte d’Ivoire, Guinea-Bissau, Mali, Niger, Senegal, Togo
XPD
Palladium (one troy ounce)
XPF
CFP franc
ng
Vietnam
French Polynesia, New Caledonia, Wallis and Futuna
432
CO2_with_index_2nd_amends.indd 432
23/05/2011 14:10
LIST OF CURRENCIES
Table A.2 (continued) XPT
Platinum (one troy ounce)
YER
Yemeni rial
Yemen
ZAR
South African rand
South Africa
ZMK
Zambian kwacha
Zambia
ZWL
Zimbabwe dollar
Zimbabwe
433
CO2_with_index_2nd_amends.indd 433
23/05/2011 14:10
CO2_with_index_2nd_amends.indd 434
23/05/2011 14:10
Bibliography
Acar, E., and A. Middleton, 2003, “Currency returns in emerging markets”, Profit & Loss, February, pp. 38–9. Acar, E., and H. Pedersen, 2001, “Which currency benchmarks for an active international investor?”, Profit & Loss, February. Acar, E., and M. Pojarliev, 2009, “Currency Management Indexes: What Do They Tell Us?”, in A. Middleton (ed.), Foreign Exchange: A Practitioner’s Approach to the Market (London: Risk Books), pp. 243–60. Acar, E., and S. Satchell, 1997, Advance Trading Rules (Oxford, England: Butterworth Heinemann). Acar, E., and R. Toffel, 2000, “Stop-loss and investment returns”, working paper, Faculty and Institute of Actuaries, Hatfield Heath. Adler, M., and B. Dumas, 1983, “International portfolio choice and corporation finance: a synthesis”, Journal of Finance, 38, pp. 925–84. Adler, M., and P. Jorion, 1992, “Universal currency hedges for global portfolios”, Journal of Portfolio Management, Summer, pp. 28–35. Adler, M., and B. Prasad, 1990, “On universal currency hedges”, Journal of Financial and Quantitative Analysis, 27, pp. 19–38. Adler, M., and D. Simon, 1986, “Exchange risk surprises in international portfolios, Journal of Portfolio Management, Winter, pp. 44–53. Aggarwal, R., 1994, “Investing in emerging markets: challenges and opportunities”, in R. Aggarwal and D. Schirm (eds), Global Portfolio Diversification (San Diego, CA: Academic Press) pp. 261–88. 435
CO2_with_index_2nd_amends.indd 435
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
Aggarwal, R., and A. L. Demaskey, 1997, “Using derivatives in major currencies for crosshedging currency risks in Asian emerging markets”, Journal of Futures Markets, 17(8), pp. 781–96. Agarwal, V., and N. Y. Naik, 2001, “Characterising hedge fund risks with buy-and-hold and option based strategies”, working paper, London Business School. AIMR, 1997a, Performance Evaluation, Benchmarks, and Attribution Analysis Charlottesville, VA: Association for Investment Management and Research. AIMR, 1997b, Managing Currency Risk, Charlottesville, VA: Association for Investment Management and Research. AIMR, 1998, “Benchmarks and performance attribution”, Subcommittee report, AIMR (August), available at http://www.aimr.com/ standards/pps/benchmark.html. AIMR, 1999, Currency Risk in Investment Portfolios, Charlottesville, VA: Association for Investment Management and Research. Alembakis, R., 2009, “Currency management forum Sydney 2009”, Global Pensions, April, pp. 18–22. Amihud, Y., 1994, “Exchange rates and the valuation of equity shares”, in Y. Amihud and R. Levich (eds), Exchange Rates and Corporate Performance, pp. 49–55 (Illinois: Irwin). Andersen, T., and T. Bollerslev, 1998, “Answering the skeptics: yes, standard volatility models do provide accurate forecasts”, International Economic Review, 39(4), pp. 885–905. Andersen, T. G., T. Bollerslev, F. X. Diebold, and P. Labys, 2001, “The distribution of exchange rate volatility”, Journal of the American Statistical Association, pp. 42–55.
436
CO2_with_index_2nd_amends.indd 436
23/05/2011 14:10
BIBLIOGRAPHY
Annaert, J., 1995, “Estimation risk and international bond portfolio selection”, Journal of Multinational Financial Management, 5(2/3), pp. 47–71. Asay, M., and C. Edelsberg, 1986, “Can a dynamic strategy replicate the return of an option?” Journal of Futures Markets, 6(1), pp. 63–70. Aston, P., 2010, “Digging Deeper into FX Transaction Costs”, Investment and Pension Asia, September–October, pp. 30 and 43. Azoff, E. M., 1994, Neural Network Time Series Forecasting of Financial Markets (New York: John Wiley & Sons). Baca, S. P., B. L. Garbe and R. A. Weiss, 2000, “The rise of sector effects in major equity markets”, Financial Analysts Journal, September/October, pp. 34–40. Bachelier, L., 1900, “Théorie de la spéculation”, Annales Scientifiques de l’Ecole Normale Supérieure, III-17, pp. 21–86; English translation in S. Haberman and T. A. Sibbett (eds), 1995, History of Actuarial Science, Vol. VII (London: Ashgate Publishing Company), pp. 15–78. Baele, L., C. Pungulescu and J.T. Horst, 2007, “Model uncertainty, financial market integration and the home bias puzzle”, Journal of International Money and Finance, 26(4), pp. 606–30. Baillie, R. T., 1996, “Long memory processes and fractional integration in econometrics”, Journal of Econometrics, 73, pp. 5–59. Balassa, Bela, 1964, “The Purchasing-Power-Parity Doctrine: A Reappraisal”, Journal of Political Economy, 72, pp. 584–96. Baldridge, J., B. Meath, and H. Myers, 2000, “Capturing alpha through active currency overlay”, Frank Russell Research Commentary (May).
437
CO2_with_index_2nd_amends.indd 437
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
Balduzzi, P., and A. Lynch, 1999, “Transaction costs and predictability: some utility cost calculations”, Journal of Financial Economics, 52, pp. 47–78. Banerjee, A. V., 1992, “A simple model of herd behaviour”, Quarterly Journal of Economics, 107, pp. 797–817. Bank for International Settlements (BIS), 2010, BIS Triennial Central Bank Survey of Foreign Exchange and Derivatives Market Activities 2010, available at http://www.bis.org. Bansal, R., and M. Dahlquist, 2000, “The forward premium puzzle: different tales from developed and emerging economies”, Journal of International Economics, 51, pp. 115–44. Barberis, N., A. Shleifer and R. Vishny, 1998, “A model of investor sentiment”, working paper, University of Chicago. Barker, W., 2007, “The global foreign exchange market: Growth and transformation”, Bank of Canada Review, Autumn, pp. 3–12. Bartov, E., and G. M. Bodnar, 1994, “Firm valuation, earnings expectations and the exchange rate exposure effect”, Journal of Finance, 49, pp. 1755–85. Basak, S., and R. Mehra, 2001, “A comparative study of portfolio insurance”, Journal of Economic Dynamics and Control, 26, pp. 1217–41. Battye, I., 2005, “Pushing for An FX Revolution”, Russell research paper. Bauer, G. H., 2001, “Conditional currency hedging and asset market shocks”, working paper, William E. Simon Graduate School of Business Administration, University of Rochester. Baxter, M., and A. Rennie, 1996, Financial Calculus: An Introduction to Derivative Pricing (Cambridge University Press).
438
CO2_with_index_2nd_amends.indd 438
23/05/2011 14:10
BIBLIOGRAPHY
Baz, J., F. Breedon, V. Naik and J. Peress, 2001, “Optimal portfolios of foreign currencies – trading on the forward bias”, Journal of Portfolio Management, 28, pp. 102–11. Beckers, C., G. Connor, and R. Curds, 1996, “National versus global influences on equity returns”, Financial Analysts Journal, March/ April, pp. 31–9. Bekaert, G., and C. R. Harvey, 2003, “Emerging market finance”, Journal of Empirical Finance, 10, pp. 3–55. Bekaert, G., and R. J. Hodrick, 1993, “On biases in the measurement of foreign exchange risk premium”, Journal of International Money and Finance, 12, pp. 115–38. Benzoni, L, P. Collin-Dufresne and R. S. Goldstein, 2005, “Can Standard Preferences Explain the Prices of Out-of-the-Money S&P 500 Put Options?”, NBER Working Paper No. W11861. Berben, R. B., and D. J. van Dijk, 1998, “Does the absence of cointegration explain the typical findings in long horizon regression?” Report 9814, Econometrics Institute, Erasmus University, Rotterdam. Berben, R. B., and D. J. van Dijk, 1999, “Unit root tests and asymmetric adjustment: a reassessment”, Research report 9902/A, Econometrics Institute, Erasmus University, Rotterdam. Berg, A., E. Borensztein and C. Pattillo, 2005, “Assessing Early Warning Systems: How Have They Worked in Practice?”, IMF Staff Papers, 52, pp. 462–502. Berkowitz, J., and L. Giorgianni, 2001, “Long-horizon exchange rate predictability?” Review of Economics and Statistics, 83, pp. 81–91. Bessembinder, H., and K. Chan, 1998, “Do the profits from technical trading rules reflect inefficiencies?” working paper, University of Utah and Hong Kong University of Science & Technology.
439
CO2_with_index_2nd_amends.indd 439
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
Bhansali, V., 2007, “Volatility and the Carry Trade”, Journal of Fixed Income, 17(3), 72-84. Billingsley, R., and D. Chance, 1996, “Benefits and limitations of diversification among commodity trading advisors”, Journal of Portfolio Management, 23, pp. 65–80. Bilson, J. F. O., 1981, “The ‘speculative efficiency’ hypothesis”, Journal of Business, 54(3), pp. 435–51. Black, F., 1989, “Universal hedging: optimizing currency risk and reward in international portfolios”, Financial Analysts Journal, July/ August, pp. 16–22. Black, F., 1990, “Equilibrium exchange rate hedging”, Journal of Finance, 45, pp. 899–908. Blake, C. R., E. J. Elton and M. J. Gruber, 1993, “The performance of bond mutual funds”, Journal of Business, 66(3), pp. 371–403. Bollerslev, T., 1986, “Generalized autoregressive conditional heteroscedasticity”, Journal of Economics, 31, pp. 307–27. Bollerslev, T., and M. Melvin, 1994, “Bid–ask spreads and volatility in the foreign exchange market”, Journal of International Economics, 36, pp. 355–72. Braccia, J. A., 1995, “An analysis of currency overlays for US pension plans”, Journal of Portfolio Management, Fall, pp. 88–93. Brandes Institute, 2007, “Currency Hedging Programs: A LongTerm Perspective”, Brandes Institute Research Paper 2007-02, April. Brennan, M., E. Schwartz and R. Lagnado, 1997, “Strategic asset allocation”, Journal of Economic Dynamics and Control, 21, pp. 1377–403. Bridgewater Associates, no date (a), “Currency observations. Developing an effective currency overlay program: the nuts and bolts”. 440
CO2_with_index_2nd_amends.indd 440
23/05/2011 14:10
BIBLIOGRAPHY
Bridgewater Associates, no date (b), “Currency observations. The dynamic hedging enigma”. Bridgewater Associates, no date (c), “Currency observations”. Bridgewater Associates, no date (d), “Currency Observations: Do Central Banks Systematically Lose Money in Currency Markets”. Brock, W., J. Lakonishok, and B. LeBaron, 1992, “Simple technical trading rules and the stochastic properties of stock returns”, Journal of Finance, 47(5), pp. 1731–64. Brunnermeier, M. K., S. Nagel and L. H. Pedersen, 2008, “Carry Trades and Currency Crashes”, NBER Working Paper Series W14473. Buraschi, A., and J. C. Jackwerth, 2001, “The price of a smile: hedging and spanning in option markets”, Review of Financial Studies, 14(2), pp. 495–527. Burik, P., and R. M. Ennis, 1990, “Foreign bonds in diversified portfolios: a limited advantage”, Financial Analysts Journal, 46(2), pp. 31–40. Burnside, A. C., M. S. Eichenbaum, I. Kleshchelski and S. Rebelo, 2006, “The returns to currency speculation”, NBER Working Paper W12489. Burnside, A. C., M.S. Eichenbaum, I. Kleshchelski and S. Rebelo, 2008, “Do Peso Problems Explain the Returns to the Carry Trade?” NBER Working Paper 14054. Burnside, C., M. Eichenbaum and S. Rebelo, 2007, “The Returns to Currency Speculation in Emerging Markets”, American Economic Review, 97(2), pp. 333–8. Campa, J. M., and P. H. K. Chang, 1998a, “The forecasting ability of correlations implied in foreign exchange options”, Journal of International Money and Finance, 17, pp. 855–80. 441
CO2_with_index_2nd_amends.indd 441
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
Campa, J. M., and P. H. K. Chang, 1998b, “Learning from the term structure of implied volatility in foreign exchange options”, in Z. Chen (ed), Currency Options and Exchange Rate Economics (Singapore: World Scientific Publishing Co.), pp. 98–115. Campa, J. M., P. H. K. Chang and J. F. Refalo, 1999, “An optionbased analysis of emerging market exchange rate expectations: Brazil’s real plan, 1994–1997”, Working Paper 6929, National Bureau of Economic Research, Chicago. Campa, J. M., P. H. K. Chang and R. Reider, 1998, “Implied exchange rate distributions: evidence from OTC option markets”, Journal of International Money and Finance, 17(1), pp. 117–60. Campbell, J. Y., K. S. Medeiros and L. M. Viceira, 2007, “Global Currency Hedging,” Harvard Business School Working Paper 07084, April. Campbell, J. Y., L. M. Viceira and J.S. White, 2003, “Foreign currency for long-term investors”, Economic Journal, 113, pp. 1–25. Canina, L., and S. Figlewski, 1993, “The informational content of implied volatility”, Review of Financial Studies, 6, pp. 659–81. Canova, F., and J. Marrinan, 1993, “Predicting excess returns in financial markets”, Working Paper 93-17, European Institute, Economics Department. Caporale, G. M., and N. Spagnolo, 2001, “Modelling East Asian exchange rates: a Markov switching approach”, working paper, South Bank University, London. Cassel, G., 1916, “The present situation on the foreign exchanges”, Economic Journal, 26, pp. 62–5. Cavaglia, S., C. Brightman and M. Aked, 2000, “The increasing importance of industry factors”, Financial Analysts Journal, 56(5), pp. 41–54. 442
CO2_with_index_2nd_amends.indd 442
23/05/2011 14:10
BIBLIOGRAPHY
Chamon, M., P. Manasse and A. Prati, 2007, “Can We Predict the Next Capital Account Crisis?”, IMF Staff Papers, 54, pp. 270–305. Chan, L. K. C., J. Karceski and J. Lakonishok, 1999, “On portfolio optimization: forecasting covariances and choosing the risk model”, Review of Financial Studies, 12(5), pp. 937–74. Chan, L., and D. Lien, 2002, “Are options redundant? Further evidence from currency futures markets”, working paper, Mount Union College/University of Texas. Chang, K. H., 2009, “Currency Hedging: A Free Lunch?” MSCI Barra research paper, April. Chase/Risk Publications, 1996, The Chase Guide to Risk Management (London: Risk Books). Chen, Y.-C., and K. Rogoff, 2003, “Commodity Currencies”, Journal of International Economics, 60, pp. 133–60. Chen, Y.-C., K. Rogoff and B. Rossi, 2010, “Can Exchange Rates Forecast Commodity Prices?,” Quarterly Journal of Economics, 125(3), pp. 1145–94. Chow, E. H., W. Y. Lee and M. E. Solt, 1997, “The economic exposure of US multinational firms”, Journal of Financial Research, 20, pp. 191–210. Chow, G., Jacquier, E. M. Kritzman and K. Lowry, 1999, “Optimal portfolios in good times and bad”, Financial Analysts Journal, 55(3), pp. 65–73. Christensen, B., and N. R. Prabhala, 1998, “The relation between implied and realized volatility”, Journal of Financial Economics, 50, pp. 125–50. Christoffersen, P., and L. Giorgianni, 1999, “Interest rate arbitrage in currency basket: forecasting weights and measuring 443
CO2_with_index_2nd_amends.indd 443
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
risk”, Working paper WP/99/16, International Monetary Fund, Washington, DC. Clarida, R. H., and M. P. Taylor, 1997, “The term structure of forward exchange premiums and the forecastability of spot exchange rates: correcting the errors”, Review of Economics and Statistics, 79, pp. 353–61. Clarida, R. H., L. Sarno, M. P. Taylor, and G. Valente, 2001, “The out-of-sample success of term structure models as exchange rate predictors: a step beyond”, working paper, Columbia University. Clarke, R. G., and M. P. Kritzman, 1996, Currency Management: Concepts and Practices (Charlottesville, VA: Association for Investment Management and Research). Clements, K. W., and R. Fry, 2008, “Commodity currencies and currency commodities”, Resources Policy, 33(2), pp. 55–73. Cline, W. R., 2008, “Estimating Consistent Fundamental Equilibrium Exchange Rates”, Peterson Institute for International Economics working paper 08-6 (July). Coakley, J., and A. M. Fuertes, 2001, “A nonlinear analysis of excess foreign exchange returns”, working paper, Manchester School. Collin-Dufresne, P., and R. S. Goldstein, 2001, “Do bonds span the fixed income markets? Theory and evidence for unspanned stochastic volatility”, working paper, Carnegie Mellon University, Department of Finance/Washington University, St Louis, and John M. Olin School of Business. Connolly, R., 1996, “Price and trading volume effects of introducing foreign exchange futures options trading”, Advances in Econometrics, 11, pp. 249–66. Conrad, J., 1989, “The price effect of option introduction”, Journal of Finance, 44(2), pp. 487–98. 444
CO2_with_index_2nd_amends.indd 444
23/05/2011 14:10
BIBLIOGRAPHY
Constantinides, G. M., 1997, “Transactions costs and the pricing of financial assets”, Multinational Finance Journal, June, pp. 93–9. Cornett, M. M., T. V. Schwarz and A. C. Szakmary, 1995, “Seasonalities and intraday return patterns in the foreign currency futures market”, Journal of Banking and Finance, 19, pp. 843–69. Covrig, V., and M. Melvin, 1998, “Asymmetric information and price discovery in the FX market: does Tokyo know more about the yen?”, working paper, Arizona State University. Cremers, M., 2000, “Stock return predictability: a Bayesian model selection perspective”, working paper, Stern School of Business, New York University. Dacorogna, M. M., U. A. Müller, C. Jost, O. V. Pictet, R. B. Olsen, and J. R. Ward, 1995, “Heterogeneous real-time trading strategies in the foreign exchange market”, European Journal of Finance, 1, pp. 383–404. Dacorogna, M. M., U. A. Müller, R. J. Nagler, R. B. Olsen, and O. V. Pictet, 1993, “A geographical model for the daily and weekly seasonal volatility in the foreign exchange market”, Journal of International Money and Finance, 12, pp. 413–38. Dales, A., and R. Meese, 2001, “Strategic Currency Hedging,” Journal of Asset Management, 2(1), pp. 9–21. De Jong, F. R., R. Mahieu, P. Schotman and I. Van Leeuwen, 1999, “Price discovery on foreign exchange markets with differentially informed traders”, discussion paper 99-32/2, Tinbergen Institute. de Kock, G., 2008, “A new fair value model for G10 currencies”, JP Morgan Global FX Strategy, September 6. de Kock, G., 2009, “Hedging and trading long-term FX with J.P. Morgan’s fair value model”, JP Morgan Global FX Strategy, April 24.
445
CO2_with_index_2nd_amends.indd 445
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
de Roon, F. A., T. E. Nijman and B. J. M. Werker, 2003, “Currency hedging for international stock portfolios: the usefulness of mean– variance analysis”, Journal of Banking and Finance, 27, pp. 327–49. De Santis, G., and G. Gerard, 1998, “How big is the premium for currency risk?” Journal of Financial Economics, 48, pp. 375–412. de Zwarty, G., T. Markwat, L. Swinkels and D. van Dijk, 2008, “The Economic Value of Fundamental and Technical Information in Emerging Currency Markets”, Erasmus Research Institute of Management Report Series, ERS-2007-096-F&A, August. Del Vecchio, F., 1999, “Currency overlay, strategies and implementation issues”, in AIMR Conference Proceedings: “Currency Risk in Investment Portfolios” (Charlottesville, VA: Association for Investment Management and Research), pp. 70–6. Del Vecchio, F., 2005, “Optimal Currency Hedging from an Australian Dollar Perspective”, presentation at Currency Management Forum, Sydney 2005. Del Vecchio, F., 2006, “Crisis Modelling of Emerging Markets: A Signal Extraction Approach”, presentation at Global Pension Currency Management Forum, Sydney. DeRosa, D., 1999, “Emerging markets and currency crises”, in AIMR conference proceedings: Currency Risk in Investment Portfolios (Charlottesville, VA: Association for Investment Management and Research), pp. 92–9. Detemple, J., and P. Jorion, 1990, “Option listing and stock returns: an empirical investigation”, Journal of Banking and Finance, September, pp. 781–801. Detzler, M. L., 1997, “The performance of global bond mutual funds”, working paper, University of Massachusetts.
446
CO2_with_index_2nd_amends.indd 446
23/05/2011 14:10
BIBLIOGRAPHY
D’Hautefort, H., and H. Xin, 2006, “The Long and Short of Currency Alpha/Beta”, Global Pensions, December. Diebold, F. X., and J. A. Lopez, 1996, Forecast Evaluation and Combination, in Maddala, G S and Rao, C R (eds), Handbook of Statistics Vol. 14, pp. 241—268: Amsterdam: North-Holland. Ding, L., and J. Hiltrop, 2010, “The electronic trading systems and bid–ask spreads in the foreign exchange market”, working paper, Macalester College and Cornerstone Research, Boston. Dissaux, Claire, 2010, “The Case for Adding Emerging Markets to Currency Overlays”, Millennium Global Research Paper, March. Doníelsson, J., and R. Payne, 2002, “Real trading patterns and prices in spot foreign exchange markets” Journal of International Money and Finance, 21, pp. 203–22. Donnelly, R., and E. Sheehy, 1996, “The share price reaction of UK exporters to exchange rate movements: an empirical study”, Journal of International Business Studies, 27, pp. 157–65. Dooley, M. P., and J. R. Shafer, 1982, “Analysis of short-run exchange rate behavior: March 1973 to November 1981”, in D. Bigman and T. Taya (eds), Exchange Rate and Trade Instability: Causes and Consequences Ballinger Publishing Co, Cambridge (Mass.), pp. 43-69. Dornbusch, R., 1976, “Expectations and exchange rate dynamics”, Journal of Political Economy, 84, pp. 1161–76. Dowd, K., 1998, Beyond Value at Risk (New York: John Wiley & Sons). Dowd, K., 1999, “A value at risk approach to risk–return analysis”, Journal of Portfolio Management, 25, pp. 60–7. Driessen, J., B. Melenberg and T. Nijman, 2000, “Common factors in international bond returns”, working paper 2000-91, Department of Economics, Tilburg University. 447
CO2_with_index_2nd_amends.indd 447
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
Duarte, A. M., and R. Rajagopal, 1999, “A scenario-based approach to optimal currency overlay”, Journal of Portfolio Management, 25(4), pp. 55–9. DuCharme, M., 2008, “The missing link in foreign exchange settlements”, Global Pensions Currency Management Supplement, January, pp. 34–5. Dumas, B., J. Fleming and R. E. Whaley, 1998, “Implied volatility functions: empirical tests”, Journal of Finance, 53(6), pp. 2059–106. Dunis, C., and B. Zhou (eds), 1998, Non-linear Modelling of High Frequency Financial Time Series (New York: John Wiley & Sons). Eaker, M., and D. Grant, 1990, “Currency hedging strategies for internationally diversified equity portfolios”, Journal of Portfolio Management, 17, pp. 30–2. Eaker, M., D. Grant and N. Woodard, 1991, “International diversification and hedging: Japanese and US perspectives”, Journal of Economics and Business, 43(4), pp. 363–74. Eaker, M., D. Grant and N. Woodard, 2000, “Realized rates of return in emerging equity markets”, Journal of Portfolio Management, 26(3), pp. 41–9. Economist, 2000, “Test-driving a new model – do currency-forecasting techniques need to be redesigned?”, March 16. Economist, 2001, Predicting the unpredictable, May 31. Ederington, L. H., and J. H. Lee, 1993, “How markets process information: news releases and volatility”, Journal of Finance, 49, pp. 1161–91. Elliott, G., and T. Ito, 1998, “Heterogeneous expectations and tests of efficiency in the yen/dollar forward exchange rate market”, working paper, University of California, San Diego. 448
CO2_with_index_2nd_amends.indd 448
23/05/2011 14:10
BIBLIOGRAPHY
Elton, E. J., M. J. Gruber and C. R. Blake, 2001, “Incentive fees and mutual funds”, working paper, New York University. Enders, W., and C. W. J. Granger, 1998, “Unit root tests and asymmetric adjustment”, Journal of Business and Economic Statistics, 16, pp. 304–11. Engel, C., 1994, “Can the Markov switching model forecast exchange rates?”, Journal of International Economics, 36, pp. 151–65. Engel, C., 1996, “The forward discount anomaly and the risk premium: a survey of recent evidence”, Journal of Empirical Finance, 3, pp. 123–92. Engel, C., and C. S. Hakkio, 1997, “The distribution of exchange rates in the EMS”, International Journal of Finance and Economics, 33, pp. 15–32. Engel, C., and J. D. Hamilton, 1990, “Long swings in the dollar: are they in the data and do markets know it?”, American Economic Review, 80, pp. 689–713. Engel, C., N. Mark and K. West, 2007, “Exchange Rate Models Are Not as Bad as You Think,” in D. Acemoglu, K. Rogoff, and M. Woodford (eds), NBER Macroeconomics Annual 2007 (Chicago, IL: University of Chicago Press). Engle, R. F., 1982, “Autoregressive conditional heteroscedasticity with estimates of the variance of U.K. inflation”, Econometrica, 50, pp. 987–1008. Engle, R., C. Hong, A. Kane and J. Noh, 1993, “Arbitrage valuation of variance forecasts using simulated options”, Advances in Futures and Options Research, 6, pp. 393–415. Errunza, V., 1997, “Research on emerging markets: past, present and future”, Emerging Markets Quarterly, Fall, pp. 5–18.
449
CO2_with_index_2nd_amends.indd 449
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
Esquivel, G., and F. Larrain, 2000, “Determinants of currency crises”, Trimestre Economico, 67, pp. 197–237. Eun, C. S., and B. G. Resnick, 1988, “Exchange rate uncertainty, forward contracts and international portfolio selection”, Journal of Finance, 43, pp. 197–215. Eun, C. S., and B. G. Resnick, 1994, “International diversification of investment portfolios: US and Japanese perspectives”, Management Science, 40, pp. 140–61. Eun, C. S., and B. G. Resnick, 1997, “International equity investment with selective hedging strategies”, Journal of International Financial Markets, Institutions & Money, 7, pp. 21–42. Eun, C. S., and B. G. Resnick, 2001, International Financial Management, 2nd edn (New York: Irwin McGraw-Hill). European Central Bank, 2003, “Review of the foreign exchange market structure”. Frankfurt am Main: ECB (March), available as PDF from http://www.ecb.int/pub/pdf/other/fxmarketstructure200303en.pdf. Evans, M. D., and R. K. Lyons, 2002, “Order flow and exchange rate dynamics”, Journal of Political Economy, 110, 170–80. Evans, M. D., and R. K. Lyons, 2005a, “Exchange Rate Fundamentals and Order Flow”, NBER Working Paper 11042. Evans, M. D., and R. K. Lyons, 2005b, “Meese-Rogoff Redux: Micro-Based Exchange Rate Forecasting”, NBER Working Paper 11042. Fama, E. F., 1965, “The behaviour of stock market prices”, Journal of Business, 38, pp. 34–105. Fama, E. F., 1984, “Forward and spot exchange rates”, Journal of Monetary Economics, 14 (November), pp. 319–38.
450
CO2_with_index_2nd_amends.indd 450
23/05/2011 14:10
BIBLIOGRAPHY
Fama, E. F., 1998, “Market efficiency, long-term returns, and behavioural finance”, Center for Research in Security Prices (CRSP), Working Paper 448, University of Chicago. Farhi, E., and X. Gabaix, 2008, “Rare disasters and exchange rates”, NBER Working Paper 13805. Faust, J., J. Rogers and J. H. Wright, 2001, “Exchange rate forecasting: the errors we’ve really made”, working paper, International Finance Division, Federal Reserve Board, Washington, DC. Fernández-Rodríguez, F., S. Sosvilla-Rivero and J. AndradaFélix, 1999, “Exchange rate forecasts with simultaneous nearestneighbour methods: evidence from the EMS”, International Journal of Forecasting, 15, pp. 383–92. Fernández-Rodríguez, F., S. Sosvilla-Rivero, and J. Andrada-Félix, 2000, “Technical analysis in foreign exchange markets: linear versus nonlinear trading rules”, WORKING paper, Fundación de Estudios de Economia Aplicada. Figlewski, S., 1997, “Forecasting volatility”, Financial Markets, Institutions and Investments, 6, pp. 2–87. Fleming, J. M., 1962, “Domestic financial policies under fixed and under floating exchange rates”, IMF staff discussion papers, 9, November, pp. 369–79. Flood, R. P., and A. K. Rose, 2002, “Uncovered interest parity in crisis”, IMF Staff Papers, 49(2). Flood, R. P., and M. P. Taylor, 1996, “Exchange Rate Economics: What’s Wrong with the Conventional Macro Approach?” in J. A. Frankel, G. Galli and A. Giovannini (eds), The Microstructure of Foreign Exchange Markets (Chicago, IL: University of Chicago Press). Flynn, A., 2005, “Currency Manager Selection and Due Diligence”, presentation at ABN-Amro Asian Currency Conference. 451
CO2_with_index_2nd_amends.indd 451
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
Folkerts-Landau, D. (ed.), 2002, “A guide to currency overlay management”, Deutsche Bank Global Markets Research Paper, February. Francis, B., I. Hasan and D. Hunter, 2002, “Emerging market liberalization and the impact on uncovered interest rate parity”, Working Paper 2002-16, Federal Reserve Bank of Atlanta. Frank, R., 2007, “Active currency management”, Mercer research paper. Frankel, J. A., 1994, “Introduction”, in J. A. Frankel (ed), The Internationalization of Equity Markets (Chicago: National Bureau of Economic Research) Frankel, J.A., and J. Poonawalaa, 2006, “The Forward Market in Emerging Currencies: Less Biased than in Major Currencies”, NBER Working Paper 12496. Frankel, J. A., and A. K. Rose, 1995, “Empirical research on nominal exchange rates”, in G. Grossman and K. Rogoff (eds), Handbook of International Economics, Vol. III, pp. 1689–729 (Amsterdam: NorthHolland). Frankel, J.A., and Wei S.-J., 2004, “Managing Macroeconomic Crises,” NBER Working Paper 10907. Froot, K. A., 1993, “Currency hedging over long horizons”, Working Paper 4355, National Bureau of Economic Research, Chicago. Froot, K. A., and J. A. Frankel, 1989, “Forward discount bias: is it an exchange risk premium?”, Quarterly Journal of Economics, February, pp. 139–61. Froot, K.A., and T. Ramadorai, 2005, “Currency Returns, Intrinsic Value, and Institutional Investor Flows”, Journal of Finance, 60, pp. 1535–66.
452
CO2_with_index_2nd_amends.indd 452
23/05/2011 14:10
BIBLIOGRAPHY
Froot, K. A., and K. Rogoff, 1995, “Perspectives on PPP and longrun real exchange rates”, in G. Grossman and K. Rogoff (eds), Handbook of International Economics, Vol. III (Amsterdam: Elsevier Science Publishers BV), pp. 1647–88. Froot, K. A., and R. H. Thaler, 1990, “Anomalies: foreign exchange”, Journal of Economic Perspectives, Summer, pp. 179–92. Fukuhara, M., 2005, “Market efficiency – can currency be managed actively?”, presentation at ABN-Amro Asian Currency Conference. Fung, W., and D. A. Hsieh, 2000, “Performance characteristics of hedge funds and commodity funds: natural vs. spurious biases”, Journal of Financial and Quantitative Analysis, 35, pp. 291–307. Fung, W., and D. A. Hsieh, 2001, “The risk in hedge fund strategies: theory and evidence from trend followers”, Review of Financial Studies, 14(2), pp. 313–41. Gabbi, G., R. Colombo, R. Branmante, M. Viola, P. Vito, and A. Tumietto, 1999, “Predicting the exchange rate: an application of econometric models, neural network and trading systems”, working paper, Euro Working Group in Financial Modelling, Vienna. Gadkari, V., and M. Spindel, 1990, “Currency hedging and international diversification – implication of a world reserve currency effect”, Journal of International Securities Markets, 4, pp. 35–42. Gallardo, P., and A. Heath, 2009, “Execution methods in foreign exchange markets”, BIS Quarterly Review, March, pp. 83–91. Garcia, C., and F. Gould, 1987, “An empirical study of portfolio insurance”, Financial Analysts Journal, 43(4), pp. 44–54. Gardner, G. W., and D. Stone, 1995, “Estimating currency hedge ratios for international portfolios”, Financial Analysts Journal, 51(6), pp. 58–64.
453
CO2_with_index_2nd_amends.indd 453
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
Gardner, G. W., and T. Wuilloud, 1995, “Currency risk in international portfolios: how satisfying is optimal hedging?”, Journal of Portfolio Management, Spring, pp. 59–67. Garza-Gómez, X., and M. Metghalchi, 2006, “Emerging Equity Markets: To Invest Or Not To Invest?”, Journal of Applied Business Research, 22(1), pp. 57–68. Gerard, B., P. Hillion and F. A. de Roon, 2002, “International portfolio diversification: industry, country, and currency effects revisited”, EFA2002 Berlin Meetings Presented Paper (March). Ghosh, S. R., and A. R. Ghosh, 2003, “Structural Vulnerabilities and Currency Crises,” IMF Staff Papers, 50, pp. 481–506. Gilmore, S.W., and F. Hayashi, 2008, “Emerging Market Currency Excess Returns”, NBER Working Paper w14528. Glen, J., and P. Jorion, 1993, “Currency hedging for international portfolios”, Journal of Finance, 48, pp. 1865–86. Goetzmann, W. N., L. Li, and K. G. Rouwenhorst, 2001, “Longterm global market correlations”, working paper, Yale School of Management, Yale University. Goldberg, P., and M. Knetter, 1997, “Goods prices and exchange rates: what have we learned?”, Journal of Economic Literature, 35 (September), pp. 1243–72. Goodhart, C., T. Ito and R. Payne, 1996, “One day in June 1993: a study of the working of the Reuters 2000–2 electronic foreign exchange trading system”, in J. A. Frankel, G. Galli and A. Giovannini (eds), The Microstructure of Foreign Exchange Markets (Chicago: University of Chicago Press), pp. 107–79. Granger, C. W. J., and Z. Ding, 1996, “Varieties of long memory models”, Journal of Econometrics, 73, pp. 61–77.
454
CO2_with_index_2nd_amends.indd 454
23/05/2011 14:10
BIBLIOGRAPHY
Grannan, E. R., and G. H. Swindle, 1996, “Minimizing transaction costs of option hedging strategies”, Mathematical Finance, 6, pp. 341–64. Griffin, J. M., and G. A. Karolyi, 1998, “Another look at the role of the industrial structure of markets for international diversification strategies”, Journal of Financial Economics, 50, pp. 351–73. Grinblatt, M., and S. Titman, 1994, “A study of monthly mutual fund returns and performance evaluation techniques”, Journal of Financial and Quantitative Analysis, 29(3), pp. 419–44. Hafeez, B., 2010, “Shedding the Label: Why FX Is No Longer Alternative”, Currency Investor Summer, pp. 10–13. Halpern, P., 1993, “Investing abroad: a review of capital market integration and manager performance”, Journal of Portfolio Management, 19(2), pp. 47–57. Hann, T. H., and E. Steurer, 1996, “Much ado about nothings? Exchange rate forecasting: neural networks vs. linear models using monthly and weekly data”, Neurocomputing, 10, pp. 323–39. Hard, S., 2005, “Currency: Emerging Markets Approach”, Global Pensions Currency Management Forum supplement. Harvey, C. R., and R. D. Huang, l991, “Volatility in the foreign currency futures market”, Review of Financial Studies, 4, pp. 543–69. Hauser, S., and A. Levy, 1991, “Optimal forward coverage of international fixed income portfolios”, Journal of Portfolio Management, 17(4), pp. 54–9. Hauser, S., M. Marcus and U. Yaari, 1994, “Investing in emerging stock markets: is it worthwhile hedging foreign exchange risk?”, Journal of Portfolio Management, 20 (Spring), pp. 76–81. Hazuka, T. B., and L. C. Huberts, 1994, “A valuation approach to currency hedging”, Financial Analysts Journal, March/April, pp. 55–9. 455
CO2_with_index_2nd_amends.indd 455
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
Henrotte, P., 1993, “Transactions costs and duplication strategies”, working paper, Stanford University. Hersey, B., and J. Minnick, 2000, “Active managers generating positive excess returns over benchmarks”, Global Pensions, February. Hersey, B., and K. Ogunc, 2000, “Designing portable alpha engines”, Investments and Pensions Europe, October. Heston, S. L., and K. G. Rouwenhorst, 1994, “Does industrial structure explain the benefits of international diversification?”, Journal of Financial Economics, 36, pp. 3–27. Hodges, S. D., and A. Neuberger, 1989, “Optimal replication of contingent claims under transaction costs”, Review of Futures Markets, 8, pp. 222–39. Hodrick, R. J., 1987, The empirical evidence on the efficiency of forward and futures foreign exchange markets, (Harwood Academic Publishers, Chur, Switzerland) Hodrick, R. J., and M. Vassalou, 2002, “Do we need multi-country models to explain exchange rate and interest rate and bond return dynamics”, Journal of Economic Dynamics and Control, 26, pp. 1275–99. Hoggard, T., A. E. Whalley and P. Wilmott, 1994, “Hedging option portfolios in the presence of transaction costs”, Advances in Futures and Options Research, 7, pp. 21–35. Hong, H., and J. C. Stein, 1997, “A unified theory of underreaction, momentum trading and overreaction in asset markets”, Working paper W6324, National Bureau of Economic Research, Chicago. Horngren, L., and A. Vredin, 1989, “Exchange risk premia in a currency basket system”, Weltwirtschaftliches Archiv, 125, pp. 311–25. Horwood, C., 2006, “How to choose the right style for alpha”, Euromoney, January. 456
CO2_with_index_2nd_amends.indd 456
23/05/2011 14:10
BIBLIOGRAPHY
Howitt, P., and R. P. McAffee, 1992, “Animal spirits”, American Economic Review, 82, pp. 493–507. Hu, M. Y., G. Zhang, C. X. Jiang, and B. E. Patuwo, 1999, “A crossvalidation analysis of neural network out-of-sample performance in exchange rate forecasting”, Decision Science, 30(1), pp. 197–216. Huang, R. D., and R. W. Masulis, 1999, “FX spreads and dealer competition across the 24-hour trading day”, Review of Financial Studies, 12, pp. 61–93. Hull, J., 2002, Options, Futures and other Derivatives, 5th edn (Upper Saddle River, NJ: Prentice Hall). Huttman, M., 2005, “Currency overlay mandates: Real issues and practical solutions”, Presentation at Global Pensions Currency Management Forum, Sydney. Ikeda, S., 1991, “Arbitrage asset pricing under exchange risk”, Journal of Finance, 46, pp. 447–55. IMF, 2010, “Global Financial Stability Report”, available at www. imf.org/external/pubs/ft/gfsr/ Isard, P., 1995, Exchange Rate Economics (Cambridge University Press). Isard, P., 2007, “Equilibrium Exchange Rates: Assessment Methodologies”, IMF Working Paper WP/07/296. Ito, T., R. K. Lyons and M. Melvin, 1998, “Is there private information in the FX market? The Tokyo experiment”, Journal of Finance, 53, pp. 1111–30. Jackwerth, J. C., and M. Rubinstein, 1996, “Recovering probability distributions from option prices”, Journal of Finance, 51(5), pp. 1611–31.
457
CO2_with_index_2nd_amends.indd 457
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
Jacquier, E., and A. J. Marcus, 2000, “Asset allocation models and market volatility”, Financial Analysts Journal, 57(2), pp. 16–30. Jagannathan, R., A. Kaplin and S. Sun, 2001, “An evaluation of multi-factor CIR models using LIBOR, swap rates, and cap and swaption prices”, Journal of Econometrics, 105(1), pp. 59–83. Johnson, S., 2009, “Investors target zero sum game”, FTfm Special Report: “Foreign exchange as an asset class”, October 6, p. 1. Jordà, Ò., and A. Taylor, 2009, “The Carry Trade and Fundamentals: Nothing to Fear but FEER Itself”, working paper, University of California, David, November. Jorion, P., 1985,” International portfolio diversification with estimation risk”, Journal of Business, 58, pp. 259–78. Jorion, P., 1990, “The exchange rate exposure of U.S. multinationals”, Journal of Business, 63(3), pp. 331–45. Jorion, P., 1994, “A mean–variance analysis of currency overlays”, Financial Analysts Journal, (May), pp. 48–56. Jorion, P., 1995, “Predicting volatility in the foreign exchange market”, Journal of Finance, 50, pp. 507–28. Jorion, P., and S. Khoury, 1995, Financial Risk Management: Domestic and International Dimensions, Blackwell: Cambridge (Blackwell). JP Morgan Fleming Asset Management, 2003, “Alternative Investment Strategies Survey” – 2002 results. Kaminsky, G. L., 2006, “Currency crises: are they all the same?”, Journal of International Money and Finance, 25, pp. 503–27. Karnosky, D. S., 1993, “Global investment in a CAPM framework”, in The CAPM Controversy: Policy and Strategy Implications for Investment Management (Charlottesville, VA: Association for Investment Management and Research), pp. 56–61. 458
CO2_with_index_2nd_amends.indd 458
23/05/2011 14:10
BIBLIOGRAPHY
Karnosky, D. S., and B. D. Singer, 1994, Global Asset Management and Performance Attribution (Charlottesville, VA: Institute of Chartered Financial Analysts). Karolyi A., and R. M. Stulz, 2002, “Are financial assets priced locally or globally?”, working paper, Ohio State University. Kedia, S., and A. Mozumdar, 1999, “Foreign currency denominated debt: an empirical examination”, working paper, Harvard University. Keynes, J. M., 1936, The General Theory of Employment, Interest and Money (Cambridge University Press). Kho, B. C., 1996, “Time-varying risk premia, volatility and technical trading rule profits: evidence from foreign currency futures markets”, Journal of Financial Economics, 41, pp. 249–90. Kidd, W. V., and B. W. Brorsen, 2004, “Why Have the Returns to Technical Analysis Decreased?” Journal of Economics and Business, 56(3), pp. 159–217. Kilian, L., 1999, “Exchange rate and monetary fundamentals: what do we learn from long-horizon regression?”, Journal of Applied Econometrics, 14, pp. 491–510. Kilian, L., and M. P. Taylor, 2001, “Why is it so difficult to beat the random walk forecast of exchange rates?”, working paper, Centre for Economic Policy Research, London. King, M. R., and C. Mallo, 2010, “A user’s guide to the Triennial Central Bank Survey of foreign exchange market activity”, BIS Quarterly Review, December, pp. 71–83. King, M. R., and D. Rime, 2010, “The $4 trillion question: what explains FX growth since the 2007 survey?” BIS Quarterly Review, December, pp. 27–42.
459
CO2_with_index_2nd_amends.indd 459
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
King, W. P., D. Klim and R. Tyagi, 2008, “Conditioning technical factors based on the currency market environment”, Global Pensions Currency Management Supplement, January, pp. 16–17. Klein, M., 1989, “Arbitrage and interest rates on currency baskets”, Weltwirtschaftliches Archiv, 125, pp. 296–310. Klement, J., and Y. Longchamp, 2010, “Managing Currency Risks for Global Families”, Journal of Wealth Management, Fall, pp. 76–87. Klepfish, E., S. M. Wright and H. Xin, 2003, “Efficient Currency Hedging”, UBS Global Equity research paper, July 8. Knez, P J., R. Litterman, and J. Scheinkman, 1994, “Explorations into factors explaining money market returns”, Journal of Finance, 49, pp. 1861–82. Kothare, S., and L. Raynor, 2010, “Are Your FX Fees Too High?”, Russell Investments Research Report, January. Kritzman, M., 1993, “The optimal currency hedging policy with biased forward rates”, Journal of Portfolio Management, Summer, pp. 94–100. Kuan, C. M., and T. Liu, 1995, “Forecasting exchange rate using feedforward and recurrent neural networks”, Journal of Applied Econometrics, 10, pp. 347–64. Kuang, P., M. Schröder and Q. Wang, 2008, “Trading rules profitability in the emerging FX market: Danger of data snooping”, accepted paper, European Financial Management Association annual meetings. Kumar, M. S., and A. Persaud, 2001, “Pure contagion and investors’ shifting risk appetite: analytical issues and empirical evidence”, Working Paper 01/134, International Monetary Fund, Washington, DC.
460
CO2_with_index_2nd_amends.indd 460
23/05/2011 14:10
BIBLIOGRAPHY
Ladekarl, J., 2005, “The World Bank Pension Plan’s experience with active currency management: A ‘fund of funds’ approach”, Presentation at Global Pensions Currency Management Forum, London. Ladekarl, J., and S. Nasypbek, 2009, “Active Currency: Worth the Risk?”, in A. Middleton (ed.), Foreign Exchange: A Practitioner’s Approach to the Market (London: Risk Books), pp. 303–38. Laïdi, A., 2008, Currency Trading and Intermarket Analysis: How to Profit from the Shifting Currents in Global Markets (New York: John Wiley & Sons) Larsen Jr, G. A., and B. G. Resnick, 2000, “The optimal construction of international diversified equity portfolios hedged against exchange rate uncertainty”, European Financial Management, 6(4), pp. 479–514. LeBaron, B., 1992, “Do moving average trading rule results imply nonlinearities in foreign exchange markets?”, working paper, Social Systems Research Institute, University of Wisconsin–Madison. LeBaron, B., 1998, “A dynamic trading strategy approach to deviations from uncovered interest parity”, working paper, Brandeis University. LeBaron, B., 1999, “Technical trading rule profitability and foreign exchange intervention”, Journal of International Economics, 49(1), pp. 125–43. Lee, C. I., K. C. Gleason and I. Mathur, 2001, “Trading rule profits in Latin American currency spot rates”, International Review of Financial Analysis, 10, pp. 135–56. Lee, C. I., and I. Mathur, 1996, “Trading rule profit in European currency spot cross-rates”, Journal of Banking and Finance, 20, pp. 949–62.
461
CO2_with_index_2nd_amends.indd 461
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
Lee, C., and S. Rahman, 1990, “Market timing, selectivity, and mutual fund performance: an empirical investigation”, Journal of Business, 63(2), pp. 261–78. Lehmann, B., and D. Modest, 1987, “Mutual fund performance evaluation: a comparison of benchmarks and benchmark comparisons”, Journal of Finance, 42(2), pp. 233–65. Leicht, G., and J. E. Tanner, 1991, “Economic forecast evaluation: Profits versus the conventional error measures”, American Economic Review, 81(3), pp. 580–90. Levich, R. M., and L. R. Thomas, 1993a, “The significance of technical trading-rule profit in the foreign exchange market: a bootstrap approach”, Journal of International Money and Finance, 12, pp. 451–74. Levich, R. M., and L. R. Thomas, 1993b, “The merits of active currency risk management: evidence from international bond portfolios”, Financial Analysts Journal, September/October, pp. 63–70. Lewis, K., 1995, “Puzzles in international financial markets”, in G. M. Grossman and K. Rogoff (eds), Handbook of International Economics, Vol. 3, pp. 1913–71 (Amsterdam: North-Holland). L’Her, F., O. Sy and M. Y. Tnani, 2002, “Country, industry, and risk factor loadings in portfolio management”, Journal of Portfolio Management, 28(4), pp. 70–9. Liang, B., 2000, “Hedge funds: the living and the dead”, Journal of Financial and Quantitative Analysis, 35, pp. 309–26. Lioui, A., 1998, “Currency risk hedging: future vs. forward”, Journal of Banking and Finance, 22, pp. 61–81. Lipscomb, L., 2005, “An overview of non-deliverable foreign exchange forward markets”, Federal Reserve Bank, New York. Lisi, F., and A. Medio, 1997, “Is a random walk the best exchange rate predictor?”, International Journal of Forecasting, 13, pp. 252–67. 462
CO2_with_index_2nd_amends.indd 462
23/05/2011 14:10
BIBLIOGRAPHY
Litterman, R., and J. Scheinkman, 1991, “Common factors affecting bond returns”, Journal of Fixed Income, 1, pp. 62–74. Lo, A. W., H. Mamaysky and J. Wang, 2001, “Foundations of technical analysis: computational algorithms, statistical inference, and empirical implementation”, Journal of Finance, 55, pp. 1705–65. Longin, F., and B. Solnik, 1995, “Is the correlation in international equity returns constant: 1960–1990?”, Journal of International Money and Finance, 14, pp. 3–26. Longstaff, F., P. Santa-Clara, and E. Schwartz, 2001, “The relative valuation of caps and swaptions: theory and empirical evidence”, Journal of Finance, 56, pp. 2067–109. Longworth, D., 1981, “Testing the efficiency of the Canadian–US exchange market under the assumption of no risk premium”, Journal of Finance, 36(1), pp. 43–9. Luca, C., 2000, Technical Analysis Applications in the Global Currency Markets, 2nd edn (Upper Saddle River, NJ: Prentice Hall). Lustig, H. N., and A. Verdelhan, 2007, “The cross-section of foreign currency risk premia and consumption growth risk”, American Economic Review, 97, pp. 89–117. Lyons, R. K., 1995, “Tests of microstructural hypotheses in the foreign exchange market”, Journal of Financial Economics, 39, pp. 321–51. Lyons, R. K., 1997, “A simultaneous trade model of the foreign exchange hot potato”, Journal of International Economics, 42, pp. 278–98. Lyons, R. K., 2002, The Microstructure Approach to Exchange Rates (Cambridge, MA: MIT Press). Malliaropulos, D., 1997, “A multivariate GARCH model of risk premiums in foreign exchange markets”, Economic Modelling, 14(1), pp. 61–79. 463
CO2_with_index_2nd_amends.indd 463
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
Malz, A., 1997, “Estimating the probability distribution of the future exchange rate from option prices”, Journal of Derivatives, Winter, pp. 20–36. Martellini, L., 2001, “Efficient option replication in the presence of transaction costs”, working paper, Marshall School of Business, University of South California. Martin, A. D., 2001, “Technical trading rules in the spot foreign exchange markets of developing countries”, Journal of Multinational Financial Management, 11, pp. 59–68. Mayhew, S., 1995, “Implied volatility”, Financial Analysts Journal, 51, pp. 8–20. Meese, R., and K. Rogoff, 1983a, “Empirical exchange rate models of the seventies: do they fit out of sample?”, Journal of International Economics, 14 (February), pp. 3–24. Meese, R., and K. Rogoff, 1983b, “The out-of-sample failure of empirical exchange rate models: sampling error or misspecification?”, in J. Frenkel (ed), Exchange Rates and International Macroeconomics (National Bureau of Economic Research and University of Chicago Press), pp. 67–105. Melvin, M., and D. Shand, 2010, “Active currency investing and performance benchmarks”, CESIFO Working Paper 3052. Menkhoff, L., and M. P. Taylor, 2007, “The obstinate passion of foreign exchange professionals: technical analysis”, Journal of Economic Literature, XLV (December), pp. 936–72. Mercer Investment Consulting, 2003, “Defined Contribution Universe Summary”, 2002 (Q2). Merk, A., 2010, “Listed Currency Vehicles – Helping Investors to Make the Right Choices”, Currency Investor, Autumn, pp. 65–70.
464
CO2_with_index_2nd_amends.indd 464
23/05/2011 14:10
BIBLIOGRAPHY
Middleton, A., 2005a, “Trading Style Analysis: A Quantitative Assessment of the Currency Industry”, Journal of Alternative Investments. 8 (2005), pp.14–28 Middleton, A (ed.), 2009, Foreign Exchange: A Practitioner’s Approach to the Market, London: Risk Books. Middleton, A., 2005b, “A Trading Style Analysis of the Parker Currency Index”, Global Money Management, February 21, pp. 8–9. Middleton, A., 2006, “A study of the currency CTA industry”, Global Pensions, Currency Management Supplement, December, pp. 38–9. Middleton, A., and S. Huang, 2007, “Currency Exposure: Does Size Really Matter?” Bank of America Global Foreign Exchange Monograph Series, No. 256. Miller, K. D., and J. J. Reuer, 1998, “Firm strategy and economic exposure to foreign exchange rate movements”, Journal of International Business Studies, 29, pp. 493–514. Mundell, R. A., 1960, “The monetary dynamics of international adjustment under fixed and flexible exchange rates”, Quarterly Journal of Economics, 74 (May), pp. 227–57. Muralidhar, A., 2001, “Why plan sponsors should hedge emerging market investments?”, FX Concepts research paper. Muralidhar, A., and H. Neelakandan, 2001, “Options to enhance currency overlay programs”, FX Concepts working paper, Alternative Investment Management Association (AIMA), London, November. Muralidhar, A., and R. J. P. van der Wouden, 2000, “Optimal ALM strategies for defined benefit pension funds”, The Journal of Risk, 2(2). Murphy, J. J., 1991, Intermarket Technical Analysis: Trading Strategies for the Global Stock, Bond, Commodity and Currency Markets (New York: John Wiley & Sons). 465
CO2_with_index_2nd_amends.indd 465
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
Murphy, J. J., 1999, Technical Analysis of the Financial Markets: A Comprehensive Guide to Trading Methods and Applications (Upper Saddle River, NJ: Prentice Hall). Muysken, B., 2005, “Currency Management: A Consultant’s View”, presentation at Global Pensions Currency Management Forum, London. National Australian Bank, 2005, “Second Australian Superannuation FX Survey” National Australian Bank, 2009, “Australian Superannuation FX Survey 2009”. Neely, C. J., 2002, “Forecasting foreign exchange volatility: is implied volatility the best we can do?”, working paper 2002-017, Federal Reserve Bank of St Louis. Neely, C. J., and P. Weller, 2001a, “Technical analysis and central bank intervention”, Journal of International Money and Finance, 20(7), pp. 949–70. Neely, C. J., and P. Weller, 2001b, “Intraday technical trading in the foreign exchange market”, working paper, Federal Reserve Bank of St Louis. Neely, C. J., P. Weller and R. Dittmar, 1997, “Is technical analysis profitable in foreign exchange markets? A genetic programming approach”, Journal of Financial and Quantitative Analysis, 32, pp. 405–26. Nesbitt, S. L., 1991, “Currency hedging rules for plan sponsors”, Financial Analysts Journal, March/April, pp. 73–81. Neuhaus, H., 1995, “The information content of derivatives for monetary policy”, discussion paper 3/95, Economic Research Group, Deutsche Bundesbank.
466
CO2_with_index_2nd_amends.indd 466
23/05/2011 14:10
BIBLIOGRAPHY
Normand, J., G. de Kock, M. Franklin-Lyons and A. Sandilya, 2010, “Managing FX Hedge Ratios: A framework for strategic and tactical decisions”, JP Morgan Global FX Strategy, May 26. Obstfeld, M., and K. Rogoff, 1996, Foundations of International Macroeconomics (Cambridge, MA: MIT Press). Okunev, J., and D. White, 2001, “Do momentum based strategies still work in foreign currency markets?”, working paper, University of New South Wales. Osler, C. L., 2000, “Support for resistance: technical analysis and intraday exchange rates”, Federal Reserve Bank of New York Policy Review, 6(2), pp. 53–68. Osler, C. L., and P. H. K. Chang, 1995, “Head and shoulder: not just a flaky pattern”, Staff Report 4, Federal Reserve Bank of New York. Pardo, R., 1992, Design, Testing, and Optimisation of Trading Systems (New York: John Wiley & Sons). Pareto Investment Management, 2009, “Emerging markets currency: better than bonds”, research paper, January. Payne, R., 2003, “Informed Trade in Spot Foreign Exchange Markets: An Empirical Investigation”, Journal of International Economics, 61, pp. 307–29. Pedersen, H., 2004, “Passive management strategies in currency overlay”, Presentation at Global Pensions Currency Forum 2004, New York. Peiers, B., 1997, “Informed traders, intervention and price leadership: a deeper view of the microstructure of the foreign exchange market”, Journal of Finance, 52, pp. 1589–614.
467
CO2_with_index_2nd_amends.indd 467
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
Perold, A., and E. Shulman, 1988, “The free lunch in currency hedging: implications for investment policy and performance standards”, Financial Analysts Journal, 44(3), pp. 45–52. Pfeuti, E., 2009, “Investors Start to query the bill”, Financial News, November 30, p. 27. Phillips, P. C. B., and A. S. Maynard, 2001, “Rethinking an old empirical puzzle: econometric evidence on the forward discount anomaly”, Journal of Applied Econometrics, 16(6), pp. 671–708. Pichardo-Allison, R., 2009, “Reducing currency volatility”, Global Pensions, April, p. 23. Pictet, O. V., M. M. Dacorogna, U. A. Müller, R. B. Olsen, and J. R. Ward, 1992, “Real-time trading models for foreign exchange rates”, Neural Network World, 6, pp. 713–44. Pikkarainen, P., 1991, “International portfolio diversification: the basket-peg regime”, Journal of International Money and Finance, 10, pp. 432–42. Pojarliev, M., 2005, “Performance of currency trading strategies in developed and emerging markets: Some striking differences”, Financial Markets and Portfolio Management 19, pp. 297–311. Pojarliev, M., 2010, “The New Normal – Defining Alpha in Currency Management”, Currency Investor Autumn, pp. 16–19. Pojarliev, M., and R. Levich, 2008a, “Do professional currency managers beat the benchmark?” Financial Analysts Journal, 64(5), September/October, pp. 18–32. Pojarliev, M., and R. Levich, 2008b, “Trades of the Living Dead: Style Differences, Style Persistence and Performance of Currency Fund Managers”, NBER Working Paper Series No. 14355.
468
CO2_with_index_2nd_amends.indd 468
23/05/2011 14:10
BIBLIOGRAPHY
Prajogi, R., A. Muralidhar and R. J. P. van der Wouden, 2000, “An asset–liability analysis of the currency decision for pension portfolios”, Institutional Investor (Derivatives Quarterly), Winter, pp. 47–56. Qi, M., and Y. Wu, 2006, “Technical trading-rule profitability, data snooping, and reality check: Evidence from the foreign exchange market”, Journal of Money, Credit and Banking, 30, pp. 2135–58. Ramaswami, M., 1993, Active Currency Management, Charlottesville, VA: Association for Investment Management and Research. Record, N., 2003, Currency Overlay (New York: John Wiley & Sons). Record, N., 2010, “An Independent Return Source”, Investment and Pension Asia, May–June, p. 14. Reiner, T. F., 2000, “Practical active currency management for global equity portfolios”, Journal of Portfolio Management, 26(4), pp. 41–8. Reuters, 2001, “New science tries to unlock secrets of the markets”, by Hament Bulsara September 12. Roberts, J., 1999, “Using double moving average rule as spot trading system”, research paper, UBS. Rogers, N., 2008, “Currency Overlay and Alpha – A detailed valueadded analysis to end Quarter 4 2007 by BNY Mellon Asset Servicing”, Presentation at Global Pensions Currency Management Forum 2008. Rogoff, K., 1996, “The purchasing power parity puzzle”, Journal of Economic Literature, 34(2), pp. 647–68. Rogoff, Kenneth, 1999, “Monetary Models of Dollar/Yen/Euro Nominal Exchange Rates: Dead or Undead?” Economic Journal, Vol. 109 (November), pp. 655–9.
469
CO2_with_index_2nd_amends.indd 469
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
Rouwenhorst, K. G., 1999, “European equity markets and the EMU”, Financial Analysts Journal, 55(3), pp. 57–64. Rozanov, A., 2004, “Currency Overlay: A New Strategy for Central Banks?”, State Street Global Advisors research paper, September. Saake, P., 2002, “Technical analysis and the effectiveness of central bank intervention”, Journal of International Money and Finance, 21, pp. 459–79. Samuelson, Paul A., 1964, “Theoretical Notes on Trade Problems,” Review of Economics and Statistics, 46, pp. 145–54. Sapp, S. G., 2002, “Price leadership in the spot foreign exchange markets”, Journal of Financial and Quantitative Analysis, 37(3), pp. 425–48. Saxena, S., and A. Villar, 2008, “Hedging instruments in emerging market economies”, in Bank of International Settlements (BIS), “Financial globalisation and emerging market capital flows”, BIS Papers No. 44, pp. 71–87. Schmittmann, J. M., 2010, “Currency Hedging for International Portfolios”, IMF Working Paper WP/10/151. Schulmeister, S., 1988, “Currency speculation and dollar fluctuations”, Quarterly Review of Banca Nazionale del Lavoro, No. 167 (December), pp. 343–65. Scott, B., 2009, “Use & Construction of Trading Models”, Presentation at FX Week China Conference, May. Secmen, A., 2009, “Is trend our friend in emerging markets?”, Citibank research paper, October. Sercu, P., 1980, “A generalization of the international asset pricing model”, Revue de l’Association Française de Finance, 1(1), pp. 91–135.
470
CO2_with_index_2nd_amends.indd 470
23/05/2011 14:10
BIBLIOGRAPHY
Sercu, P., and R. Vanpée, 2007, “Home bias in international equity portfolios: a review”, Leuven School of Business and Economics Working Paper. Shleifer, A., and L. H. Summers, 1990, “The noise trader approach to finance”, Journal of Economic Perspectives, 4, pp. 19–33. Simotas, P., 2002, “Bring prime brokerage to currency overlay”, FX Week, December 23, 2002. So, J. C., 1987, “The distribution of foreign exchange price changes: trading day effects and risk measurement – a comment”, Journal of Finance, 42, pp. 181–8. Solnik, B., 1974, “An equilibrium model of the international capital market”, Journal of Economic Theory, 8, pp. 500–24. Solnik, B., 1983b, “International arbitrage pricing theory”, Journal of Finance, 38(2), pp. 449–57. Solnik, B., 1993, “The performance of international asset allocation strategies using conditioning information”, Journal of Empirical Finance, 1, pp. 33–55. Solnik, B., C. Beucrelle and Y. Le Fur, 1996, “International market correlation and volatility”, Financial Analysts Journal, September/ October, pp. 17–34. Sorensen, E. H., J. J. Mezrich and D. N. Thadani, 1993, “Currency hedging through portfolio optimisation”, Journal of Portfolio Management, Spring, pp. 78–85. Sorescu, S., 2000, “The effect of options on stock prices: 1973–1995”, Journal of Finance, 55, pp. 487–514. Stannard, J. C., et al, 1998, “Benchmarks and Performance Attribution Subcommittee Report”, Association for Investment Management and Research, August (available at URL: www.aimr.com/ standards/pps/benchmark.html). 471
CO2_with_index_2nd_amends.indd 471
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
Stehle, R., 1977, “An empirical test of the alternative hypotheses of national and international pricing of risky assets”, Journal of Finance, 32, pp. 493–502. Stulz, R. M., 1984, “Pricing capital assets in an international setting: an introduction”, Journal of International Business Studies, Winter, pp. 55–72. Stulz, R. M., 1995, “International portfolio choice and asset pricing: an integrative survey”, in V. Maksimovic and W. Ziemba (eds), The Handbook of Modern Finance (Amsterdam: North-Holland) pp. 201–23. Sullivan, R., A. Timmermann and H. White, 1998, “Data-snooping, technical trading rule performance, and the bootstrap”, working paper, University of California at San Diego. Surajaras, P., and R. J. Sweeney, 1992, Profit-making Speculation in Foreign Exchange Markets (Boulder, CO: Westview Press). Sweeney, J., 1986, “Beating the foreign exchange market”, Journal of Finance, 41, pp. 163–82. Szakmary, A. C., and I. Mathur, 1997, “Central bank intervention and trading rule profits in foreign exchange markets”, Journal of International Money and Finance, 16, pp. 513–35. Taylor, A. M., 2001, “Potential pitfalls for the purchasing-powerparity puzzle? Sampling and specification biases in mean-reversion tests of the law of one price”, Econometrica, 69, pp. 473–98. Taylor, A., and A. Farstrup, 2006, “Active Currency Management: Arguments, Considerations, and Performance for Institutional Investors”, CRA RogersCasey International Equity Research Paper, June. Taylor, M. P., 1995, “The economics of exchange rates”, Journal of Economic Literature 83, pp. 13–47. 472
CO2_with_index_2nd_amends.indd 472
23/05/2011 14:10
BIBLIOGRAPHY
Taylor, M. P., and H. Allen, 1992, “The use of technical analysis in the foreign exchange market”, Journal of International Money and Finance, 11, pp. 304–14. Taylor, S. J., 1986, Modelling Financial Time Series (Chichester, England: John Wiley & Sons). Taylor, S. J., 1994, “Trading futures using a channel rule: a study of the predictive power of technical analysis with currency examples”, Journal of Futures Markets, 14, pp. 215–35. Thomas III, L. R., 1989, “The performance of currency-hedged foreign bonds”, in L. R. Thomas III (ed), The Currency Hedging Debate (London: IFR Publishing) (first published in Financial Analysts Journal, May/June, pp. 25–31). Thomas III, L. R. (ed), 1990, The Currency Hedging Debate (London: IFR Publishing). Urrutia, J., and J. Vu, 2001, “The impact of primes and scores on the price, volatility and trading volume of the underlying stocks”, Journal of Financial Practice and Education, Fall. VanderLinden, D., C. X. Jiang and M. Hu, 2002, “Conditional hedging and portfolio performance”, Financial Analysts Journal, 58(4), pp. 72–82. Vitale, P., 2006, “A Guided Tour of the Market Microstructure Approach to Exchange Rate Determination”, working paper, D’Annunzio University. Waki, Natsuko, 2002, “FX fixing proves popular among nervous investors”, Reuters, November 1. Waring, M. B., and L. B. Siegel, 2005, “Debunking some myths of active management”, Journal of Investing, 14(2), pp. 20–8.
473
CO2_with_index_2nd_amends.indd 473
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
Wasserfallen, W.,, 1989, “Flexible exchange rates: a closer look”, Journal of Monetary Economics, 23 (May), pp. 511–21. Weigend, A. S., D. E. Rumelhart and B. A. Huberman, 1991, “Generalisation by weight elimination with application to forecasting”, Advances in Neural Information Processing Systems, 3, pp. 875–82. Weinberg, S. A., 2001, “Interpreting the volatility smile: an examination of the information content of option prices”, International Finance Discussion Paper 706, Federal Reserve Board of Governors. West, K. D., and D. Cho, 1995, “The predictive ability of several models of exchange rate volatility”, Journal of Econometrics, 69(2), pp. 367–91. Whalley, A. E., and P. Wilmott, 1993, “An asymptotic analysis of the Davis, Panas & Zariphopoulou model for option pricing with transaction costs”, OCIAM working paper, Mathematical Institute, Oxford. Williams, S., 2005, “A New Signal for Model Trading: Using FX market correlations to boost your alpha”, Presentation at Global Pensions Currency Management Forum, London. Williamson, J., 1983, “The Exchange Rate System”, Policy Analyses in International Economics, 5. Xin, H., 1999, “Using risk reversals to predict spot movement”, research paper, UBS. Xin, H., 2003, “Currency Overlay”, Currency Overlay: A Practical Guide (London: Risk Waters Group). Xin, H., 2010, “Toolkit for Forecasting FX: Past, Present and Future”, presentation at FX Week Asia Conference, October. Xin, H., and H. Rondouin, 2009, “From Active Overlay to Passive Overlay Plus Pure Alpha”, in A. Middleton (ed.), Foreign Exchange: A Practitioner’s Approach to the Market (London: Risk Books), pp. 91–105. 474
CO2_with_index_2nd_amends.indd 474
23/05/2011 14:10
BIBLIOGRAPHY
Zhang, G., and M. Y. Hu, 1998, “Neural network forecasting of the British pound/US dollar exchange rate”, International Journal of Management Science, 26(4), 495–506. Zhu, Y., and R. Karee, 1988, “Performance of portfolio insurance strategies”, Journal of Portfolio Management, Spring, pp. 48–54. Zitzewitz, E. W., 2001, “Measuring exaggeration by analysts”, working paper, Stanford University.///
475
CO2_with_index_2nd_amends.indd 475
23/05/2011 14:10
CO2_with_index_2nd_amends.indd 476
23/05/2011 14:10
Index
(page numbers in italic type denote tables and figures)
A absolute currency returns 15–18 Active Currency Management xxix active overlay 190–3 and benchmark hedge ratio 190 currency pure alpha, trading styles for 104–9 classification of 106 fundamental-discretionary 108 fundamental-quantitative 106–7 technical-discretionary 107–8 technical-quantitative 108–9 and reporting 192–3 and symmetry of constraints 190–2 see also passive overlay actual volatility 72, 360, 366, 368 does implied volatility overpredict? 362–4 adjustment style 324–8 administration cost 182, 275 alpha–beta separation 80, 331–43 active versus passive plus pure alpha, summary of 341–2 best implementation methods for 342–3 for currencies 331–6 alpha as beta and vice versa 332–3 definitions of 331–2 historical background of 333–4 industry trend 335–6 and overlay 331–43 and rationale behind passive overlay plus pure alpha 336–42 and active, comparing 338–9 adding pure alpha 340–1
key benefits of passive, plus a currency-pure-alpha programme 342 Aon Consulting 124 arbitrage 70–1 Arch (autoregressive and conditional heteroscedastic) 365, 366 asset–liability optimisation 293–5 Association for Investment Management and Research (AIMR) 302, 303, 307 asymmetrical benchmarks 191 at-the money-forward (ATMF) 363, 366 attribution process 314 Autobahn 397 autocorrelation (momentum, trends and countertrends) 50 B Bank for International Settlements (BIS) 8, 12, 376, 395, 396 Bank of England 2, 188, 365 Bank of New York (BNY) Mellon 123, 127 Barclay Capitals 121 BarclayHedge Currency Trader Index 126 BARX 397 basket peg 387–8 Bayesian updating 235–6 Bear Stearns 399 benchmark hedge ratios 172–4, 190, 416–20; see also hedge ratios benchmarks: customised, and risk separation and risk responsibility 304–6 illustration of 305 design of, and performance
477
CO2_with_index_2nd_amends.indd 477
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
measurement 301–29 adjustment style 324–8 and currency returns, dissecting 306–7; see also currency returns currency-surprise approach 307–12 excess-return approach 312–15 and hedging actual exposure 324 and hedging benchmark weighting 324–7 and hedging to benchmark neutral 328 some practical considerations concerning 328 holding-based 315–24 performance-based benchmarks, problems with 135–6 setting, qualities of 302–3 bid–offer spread 8, 179–81 Big Mac Index 203–4 binomial representation 347, 349 Black–Scholes–Merton option-pricing theory 345, 355, 359–60 Black’s universal hedge ratio 283–4 BNP Paribas 121, 123 Bollinger bands 207, 230 bond prices and currency fluctuations 38–9 bootstrap methods 228, 242, 288 Brandes Institute 54, 173, 178 Bretton Woods conference 2 Bridgewater Associates 112 burgernomics 204 C Callan Associates 54 CalPERS 97, 146 CalSTRS 146 Cambridge Associates 124 canonical model of exchange rates 211 capital asset pricing model (CAPM) 47, 52, 172, 267 and currencies 268–74 background to 268–70 and foreign-exchange hedging and
overlay 274–6 international (ICAPM) 267, 269–74 basic models 270–4 optimisation of, for maximum risk adjusted returns 173 career forecasters 203, 257 carry and trend 218–31 and carry-trade improvements 222–3 and carry trade’s long history 218–20 and effectiveness 220–2, 226–9 and forward bias 218–24 and momentum 224–31; see also main entry and value in EM currencies 381–5 cash management 96, 181, 183, 410 cashflow 17, 19, 84, 159, 190, 328 impact of 174, 181 “problem of hedging” 79 central bank intervention 258–9 Centre for International Banking Economics and Finance index (AFX Currency index) 250–1 CISDM CTA Currency Index 126 Citibank 121 Citigroup 123 CLS Bank 404–5 commodity trading advisers 127, 196 competency 94–6, 160, 175, 406, 414–15 constraints: risk 166–72 different restrictions for different types of currency programme 171–2 leverage limit 168–9 “no net short sale”, “no leverage” and other position limits 166–8 stop-loss and maximum drawdown limits 170 tolerance range around hedge ratio 170–1 value-at-risk limit 169–70 volatility and tracking-error limit 168
478
CO2_with_index_2nd_amends.indd 478
23/05/2011 14:10
INDEX
symmetry of 190–2 continuous-time finance 345, 346, 349, 355 continuous linked settlement 404–5 copulas 240 correlation forecasting 365–6 correlation triangle 71–2 counterparty banks and prime brokers 121–3 counterparty credit risk: and liquidity and e-commerce 393– 412 passim Credit Suisse 121 Credit Suisse (CS) index 250–1 cross-border merger-and-acquisition and foreign-exchange announcements 259–63 cross-hedging 186–7 currencies: and diversification benefit, hedging reduces 45–6 efficiency of market in 47–8 and foreign exchange, as not a separate asset class 43–5 hedging of, see currency hedging; hedging list of 425–33 movements in, as unpredictable random walks 42–3 part of global diversification 45–6 returns from, see currency returns risk premium not attracted by 46–7 risks concerning, see currency risks special nature of 40–52 and autocorrelation 50 and forward bias 48–50 trading in, a “zero-sum game” 41–2 currency basket hedging 295–8 basket to base currency 295–7 basket to basket 297 see also hedging currency beta indexes 247–51 Centre for International Banking Economics and Finance (AFX Currency index) 250–1
Credit Suisse (CS) 250–1 definition and varieties of 247–8 Deutsche Bank (DB) 250 examples and performance of 249–50 impact of 248–9 potential problems with 250–1 currency fixing arrangements 188–9 currency, emerging-market 373–91 carry and trend in 381–5 exposures, whether to hedge 376–8 hedging issues for investors based on 380–1 more recent interest in 375–6 options to trade 385–6 capturing carry by buying call options on 386 overview of 385 selling and selling volatility 386 other issues related to 386–90 basket peg 387–8 cross-correlation 386–7 indexes of risk and risk appetite 388–90 and selective hedging 379–80 special characteristics of 374–5 trend-following in 383–4 using, to forecast spot 386 value-based trading in 385 currency hedging 23, 29, 55–69 passim, 90 debate 2, 40–1 impact 46 survey of studies on 59–69 see also currencies; hedging currency management: styles 201–65 passim; see also currency management styles and sector-based investment 298–9 Currency Hedging Debate, The xxx Currency Management: Concepts and Practices xxx currency, management styles 201–65 passim and carry and trend 218–31 and carry trade’s long history 218–20
479
CO2_with_index_2nd_amends.indd 479
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
and carry-trade improvements 222–3 and carry-trade improvements 222–3 and effectiveness 220–2, 226–9 and forward bias 218–24 and momentum 224–31; see also main entry and currency beta indexes 247–51 Centre for International Banking Economics and Finance (AFX Currency index) 250–1 Credit Suisse (CS) 250–1 definition and varieties of 247–8 Deutsche Bank (DB) 250 examples and performance of 249–50 impact of 248–9 potential problems with 250–1 and foreign-exchange determination, microeconomic models of 210–16; see also foreign-exchange determination and quantitative forecast models, other types of 231–46 Bayesian updating 235–6 flow models 239 idiosyncratic 246 market microstructure and high frequency data 236–9 Markov-switching 233–5 other data-mining techniques 239–46 principal component analysis, classification and tree methods 236 quantitative foreign-exchange modelling in general 231–3 and technical analysis 216–18 currency overlay: and alpha–beta separation 331–43 active versus passive plus pure alpha, summary of 341–2 best implementation methods for 342–3
for currencies 331–6 and rationale behind passive overlay plus pure alpha 336–42 and alpha–beta separation 331–43 and CAPM and foreign-exchange hedging 274–6 checklist for 420 evaluating before and after 423 forming a strategic view on currency returns 421 forming a strategic view on currency risk 420–1 more choices 422–3 reviewing managers’ core competency 422 selecting a passive hedge ratio 421 selecting a trading style 422 consultants 124 cost of 144–8 forward premium/discount 148 management and performance fees 144–6 opportunity costs 147–8 transaction costs 146–7 custodians 123–4 defining a 98 definitions and merits of 86–97 working definition 88–9 difference between hedging and 90 different types of, explained 91–3 active 91 enhanced 91 passive 91 passive/active into currency basket 92 pure-alpha 91–2 tactical 92 and dynamic hedging and options 345–71 option-pricing theory 345–59 and skews and smiles 359–61 and transaction costs and option replication 349–50
480
CO2_with_index_2nd_amends.indd 480
23/05/2011 14:10
INDEX
and dynamic hedging and trend following 353–4 employing, advantages of 96–7 employing, disadvantages of 97 end-users of 120–1 evaluating 149–50 organisation 149–50 processes 150 track record 149 explained 69–70 and foreign assets, portion of portfolio allocated to 172 internal and external programme, choice between 160 passive and currency-pure-alpha, running together 159–60 using several managers 160–1 hedging approach to, why used 350–3 laid on different asset classes 98–100 main parties involved in 109–24 managers 109–21; see also currency-overlay manager management of 77–155 competency-based approach to 94–6; see also competency outsourcing 140–50; see also outsourcing currency management practical choices concerning 155–200 and programmes, historical performance of 124–40; see also currency-overlay programmes and risk separation/responsibility 93–4 and risks, approaches to 81–6; see also under currency risk manager of, see currency-overlay manager market, main players in 151–2 origins and development of 78–81 practical choices concerning management of 155–200 active and currency-pure-alpha, choice between 158–9
active and passive, choice between 156–7 for active 190–3; see also active overlay and hedging horizon 177–8 hedging benchmark weights or exposures 175–7 instruments 161–6 for overall 156–172 for passive 172–90; see also passive overlay for pure alpha 193–6; see also pure alpha risk constraints 166; see also main entry and prime brokerage 196–9; see also prime brokerage process of, in practice 97–104 and different parties, information flow between 101–2 programmes, historical performance of 124–40; see also currency overlay programmes and risk management 52–70 continued interest in 52–5 and growing global market and its consequences for investors 55–6 and volatility and correlation forecasting 365–6; see also volatility currency-overlay managers: AUM and AUME figures presented by 115–17 daily flowchart for 102–4 decision making within 104 early performance studies concerning 127–8 evaluating 149–50 organisation 149–50 processes 150 track record 149 internal and external 109–10 investment process for 105
481
CO2_with_index_2nd_amends.indd 481
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
major players 112 overview of 110–12 recent changes 118–19 return-enhancement versus risk control objectives of 114–15 services provided by 113–14 trading styles of 117–18 using several 160–1 currency-overlay programmes 124–40 foreign and foreign-exchange-only performance databases and indexes 126–7 performance of, new studies on 136–40 see also currency overlay: management of currency pure alpha: and active overlay, trading styles for 104–9 choices for 193–6 funded or unfunded 193–4 and listed vehicles 194 and reporting 195–6 and setting a risk budget 195 and target information ratios and target returns 195 classification of 106 currency returns: absolute, when underlying investment return is certain 15–18 suitable return measure, selecting 18–19 absolute, when underlying investment return is uncertain 19–21 defining 13–25 dissecting 306–7 ex ante versus ex post 23–4 nominal versus real 24–5 relative 22–3 unhedged, hedged and local currency 25 currency risk: approaches to managing 81–6 currency-pure-alpha 86
doing nothing 81–2 and hedging, active v. passive v. dynamic 82–6 and correlation 35–6 descriptive tour of 26–40 fluctuations, and bond prices 38–9 fluctuations, and stock prices 36–8 hedging of, questions and answers concerning 56–9 hedging of, survey concerning 59–69 separation and responsibility 93–4 short-term versus long-term contribution 26–34 and statistics, word of caution concerning 39–40 Currency Risk in Investment Portfolios xxix currency, role of, and global optimisation 276–80 currency separation 414 currency surprise approach 302, 307–12, 319 Currenex 397 custodian banks and eFX 411–12; see also e-commerce custodians 123–4 customised benchmarks: risk separation and risk responsibility 304–6; see also benchmarks D Daiwa Fund Consulting 124 data-mining: pitfalls of 255 techniques for 239–46 defining currency returns 13–25 derivatives 162 OTC and exchange-traded, recent trends in 12–13 Derivatives Market Activities 8 Deutsche Bank 121 Deutsche Bank (DB) index 250 Deutsche Bank FX Select platform 126 dissecting currency returns 306–7 diversification:
482
CO2_with_index_2nd_amends.indd 482
23/05/2011 14:10
INDEX
global, currency as 45–6 hedging reduces benefit of 45–6 Dornbusch model 211, 212 E EBS 397 e-commerce: and continuous linked settlement 404–5 and currency management in practice 410–11 and custodian banks 411–12 general observations on 403–4 and liquidity and counterparty risk 393–412 and liquidity, impact on 396–8 and multibank platforms 405–8 and single-bank platforms 408–10 and straight-through processing 404 economic fundamental models 213–14 Economist 203, 211, 231, 232 ECU 3 eFX, see e-commerce emerging markets (EMs): background to 373–4 currencies 373–91 carry and trend in 381–5 exposures, whether to hedge 376–8 hedging issues for investors based on 380–1 more recent interest in 375–6 options to trade 385–6 other issues related to 386–90 and selective hedging 379–80 special characteristics of 374–5 trend-following in 383–4 value-based trading in 385 hedging debate in 374–81 equilibrium real exchange rate (ERER) 209 European Monetary System (EMS) 3 ex post data 228, 241, 366 excess-return approach 312–15
exchange-traded derivatives (ETDs) 12–13, 162, 163 exotic options 229, 361 exponential smooth threshold autoregressive (ESTAR) model 215 extreme-value theory (EVT) 240 F filter rules 271, 225, 227 fixing arrangements 188–9 floating-rate system 3, 34 forecast models, quantitative 231–46 Bayesian updating 235–6 flow models 239 idiosyncratic 246 market microstructure and high frequency data 236–9 Markov-switching 233–5 other data-mining techniques 239–46 genetic programming 240–2 nearest-neighbour methods 243–4 neural-network algorithms 244–6 and seasonality 242–3 principal component analysis, classification and tree methods 236 and quantitative foreign-exchange modelling in general 231–3 forecast models, relevance of, in portfolio-management context 257–8 forecast, forward and no-arbitrage price 72–4 forecasting in the financial market 255–6 foreign exchange (FX): electronic, see e-commerce not a separate asset class 43–5 old and new market structures 7 Parker index in, analysis of 128–35 research, and choice of currency management styles 201–65; see also currency management styles and risk management and currency overlay 52–70 continued interest in 52–5
483
CO2_with_index_2nd_amends.indd 483
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
and growing global market and its consequences for investors 55–6 foreign-exchange and cross-border merger-and-acquisition announcements 259–63 foreign-exchange derivatives 4, 12, 13, 141, 357 foreign-exchange determination: microeconomic models of 210–16 brief history of 210–12 and commodity currencies 214 practical use of, and new angles of research 214–16 stylised representation of 212 validity of, empirical evidence on 213 and technical analysis 216–18 foreign-exchange e-commerce, see e-commerce foreign-exchange liquidity, see liquidity foreign-exchange market: daily turnover of 9 changing structure of 7 and recent trends in OTC and exchange-traded currency derivatives 12–13 stylised facts about 5–13 participants in 5–6 types of trade, transparency and transaction cost in 6–7 size of, major currencies in and concentration of 8–12 forward bias 29, 48–50, 71, 218–24 and carry trade’s effectiveness 220–2 and carry-trade improvements 222–3 and carry trade’s long history 218–20 reason for existence of 223–4 forward rate agreements (FRAs) 72 forward volatility agreements (FVAs) 366–7 Frank Russell Company 127, 128 frequency and tolerance band, rebalancing 182
Frontier Investment Consulting 124 FTSE Currency Forward Rate Bias Index 250–1 fundamental equilibrium exchange rate (FEER) 209, 385 fundamental models 210–16 funded pure alpha 193–4 FX Concepts 112 FX Connect 397 FX research 201–65 passim and carry and trend 218–31 and carry trade’s long history 218–20 and carry-trade improvements 222–3 and effectiveness 220–2, 226–9 and forward bias 218–24 and momentum 224–31; see also main entry and currency beta indexes 247–51 definition and varieties of 247–8 examples and performance of 249–50 impact of 248–9 potential problems with 250–1 and foreign-exchange determination, microeconomic models of 210–16; see also foreign-exchange determination putting into action 251–7 and data mining, pitfalls of 255; see also data-mining practicalities 255–7 and recent performance and market trend 252–4 toolkit for 251–2 and quantitative forecast models, other types of 231–46 Bayesian updating 235–6 flow models 239 idiosyncratic 246 market microstructure and high frequency data 236–9 Markov-switching 233–5
484
CO2_with_index_2nd_amends.indd 484
23/05/2011 14:10
INDEX
other data-mining techniques 239–46 principal component analysis, classification and tree methods 236 quantitative foreign-exchange modelling in general 231–3 and technical analysis 216–18 FXall 397, 407–8 FXCM 397 FXTrader 397 G gamma scalping 367–8 Gaudre, Sebastien 207 generalised autoregressive conditional heteroscedasticity (Garch) 365 genetic programming 232, 240–2 global diversification, and currency 45–6 Global Harvest Index 382 global market and its consequences for investors 55–6 global optimisation and the role of currency 276–80 global tactical asset allocation (GTAA) 80, 89, 119 Goldman Sachs 112, 121 “Great Moderation” 4, 399 Greeks 369 H hedge ratios: and asset–liability optimisation 293–5 benchmark 172–4, 190, 416–20 Black’s universal 283–4 and currency returns, forming a strategic view on 421 dynamic 282–3 factors affecting choice of 290–3 minimum-variance 284–5 optimisation and 267–300 and CAPM and currencies 268–74 empirical evidence 285–7
and parameter uncertainty and structural models 287–9 and parameter uncertainty and structural models 287–9 deciding on and selecting a passive 421 tolerance range around 170–1 unitary 281–2 and variance–covariance matrixes, instability of 289–90 hedging: active v. passive v. dynamic 82–6 active: cons 84–5 active: pros 84 dynamic: cons 85–6 dynamic: pros 85 passive: cons 83–4 passive: pros 83 to benchmark neutral 177, 328 benchmark weights or exposures 175–7 and CAPM 274–6 cash account 183–6 cross- 186–7 and currency risk, questions and answers concerning 56–9 currency basket 295–8 basket to base currency 295–7 basket to basket 297 currency, difference between overlay and 90 currency, survey of studies on 59–69 and diversification benefit 46 dynamic, option-pricing theory and 346–59 intuitive points from 348–9 dynamic, and overlay 345–71 and skews and smiles 359–61 and transaction costs and option replication 349–50 dynamic, to protect a “floor” 354–5 dynamic, and trend-following 353–4 dynamic, why used as approach to currency overlay 350–3
485
CO2_with_index_2nd_amends.indd 485
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
and EM exposures 376–8 actual exposure 324 benchmark weighting 324–7 horizon 177–8 and issues for investors based on EM currencies 380–1 option price determined through 346–8 proxy 187–8 selective 379–80 and volatility and correlation forecasting 365–6; see also volatility Hersey and Minnick 127 Hewitt Associates 124 high-frequency and algorithmic trading 399 holding-based benchmarks 315–24 Hotspot FX 397 HSBC 121, 123 I implied volatility 72, 360, 366, 368 and prediction of actual volatility 362–4 indexes of risk, and risk appetite 388–90 information flow 101–2, 110, 126, 394, 410 instrument tenor 177, 178–82 instruments for currency-overlay management 161–6 currency options 165–6 derivatives 162 foreign-exchange swaps 163–4 forwards versus futures 162 innovative, other types of 166 non-deliverable forwards 164–5 Intech 124 interest-rate arbitrage, uncovered 70–1 interest-rate spread 224 international capital asset pricing model (ICAPM) 177, 267, 269–74 basic models 270–4 inflation 273–4 single-factor 270–1
Solnik–Sercu currency-hedging 271–3 empirical evidence on 274 International Financial Management xxx International Monetary Fund (IMF) 2 International Swaps and Derivatives Association (ISDA) agreements 101, 102, 141 and exceptions 143–4 J Jana Investment Advisers 124 Japanese yen 387 Jensen’s inequality 284 JP Morgan 112, 209 JP Morgan Fleming 53 K Korean won 386–7 L law of one price 51, 203, 203–4 layered forward contracts 185 learning models 215, 235 Lehman’s 399 liquidity: and counterparty risk and e-commerce 393–412 electronic dealing’s impact on 396–8 other factors impacting on (2000– 2010) 398–401 theory and reality of 394–401 Liquidity, Credit and Volatility Index 389 M Maastricht Treaty 3 macroeconomics, failure of 106–7 management and performance fees 144–6 Managing Currency Risk xxix market microstructure and high frequency data 236–9 Markov-switching models 233–5 mean reversion 34, 87, 173, 207 countertrending and 230
486
CO2_with_index_2nd_amends.indd 486
23/05/2011 14:10
INDEX
mean–variance portfolio optimisation 44, 267, 285 Mellon Analytical Solutions 127 Menchen, H. L. 212–13 Mercer Investment Consulting 53, 127 Mercer, Watson Wyatt 124 Millennium Asset Management 112 Millennium Partners 215 minimum-variance currency hedge 272, 275 modern currencies, brief history of 2–5 momentum 50 countertrending and mean-reversion 230–1 and effectiveness 226–9 existence of 231 and filter rules 225 and moving-average rules 225–6 and stop-loss and take-profit 229–30 variations on a theme of 224–31 Morgan Stanley 121 moving-average rules 225–6 MSCI EAFE Index 26, 167, 168, 289, 332 MSCI Index 239 multibank platforms 405–8 and e-commerce 405–6; see also e-commerce Mundell–Fleming framework 311, 212 N National Association of Pension Funds 55 National Australian Bank 53–4 nearest-neighbour methods 243–4 neural-network algorithms 218, 244–6 nominal versus real currency return 24–5 nominal-return-based approach 34 non-deliverable forwards (NDFs) 164–5, 189 Northern Trust 123 O Olsen Data 238 “optimal band” approach 349 optimal currency-hedge ratio 281; see
also hedge ratios “optimal time interval” 349 optimisation 255, 350, 415, 416–17 asset–liability 293–5 and CAPM, for maximum risk adjusted returns 173 global, and role of currency 276–80 and hedge ratios 267–300 and CAPM and currencies 268–74 empirical evidence 285–7 and parameter uncertainty and structural models 287–9 mean–variance portfolio 44, 267, 285 portfolio, for minimum portfolio variance 173 option premiums 85, 185, 322, 353, 358, 359 option pricing theory of 345–59 options: currency, as instruments for overlay management 165–6 for currency-overlay programmes 358–9 and dynamic hedging and overlay 345–71 approach to, why used 350–3 intuitive points from 348–9 to protect a “floor” 354–5 and transaction costs and option replication 349–50 and dynamic hedging and trend following 353–5 market, trading opportunities in 361–9 overview of 361–2 pricing theory of 345–59 purchased 185 redundancy of 355–8 replication of 83, 92–3, 114, 281, 282, 338, 345, 353–4 and transaction costs 349–50 and skews and smiles 359–61 to trade EM currencies 385 Ornstein–Uhlenbeck process 230
487
CO2_with_index_2nd_amends.indd 487
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
outsourcing currency management 140–50 legal documents for 141–4 investment-management agreement 141–3 ISDA agreement 143–4 prime-brokerage agreement 144 over-the-counter (OTC) 8, 121, 162, 237, 357, 369, 393 Overlay Asset Management 112 P parameter uncertainty 59, 282, 287–9 Pareto 112 Parker FX indexes 126, 253 analysis of 128–35 Parker Global Strategies (PGS) 128–9, 136 passive overlay 92, 114, 145, 155, 172–90 active versus, plus pure alpha, summary of 341–2 and benchmark hedge ratio 172–4 and bonds and equities, same or different benchmark hedge ratio for 175 choices for 172–90 and currency-fixing arrangements 188–9 and currency-management, active versus 156–7 with currency-pure-alpha 159–60 and hedging benchmark weights or actual exposures 175 actual exposure 176 to benchmark neutral 177 benchmark weighting 176–7 see also hedging and hedging cash account 183–6 and instrument tenor 178–82 key benefits of, plus a currency pure-alpha programme 342 and pure alpha, rationale behind 336–42 and active, comparing 338–9 adding pure alpha 340–1
and rebalancing frequency and tolerance band 182 reporting 189–90 and reporting 189 and rollover date 186 see also active overlay pegging 387–8 Pension Fund Association 124 pension funds 24, 46, 54–5 CalPERS 97, 146 CalSTRS 146 as currency-overlay end-users 120 performance and management fees 144–6 performance measurement and benchmark design 301–29 adjustment style 324–8 and hedging actual exposure 324 and hedging benchmark weighting 324–7 and holding-based benchmarks 315–24 currency-surprise approach 307–12 and dissecting currency returns 306–7; see also currency returns excess-return approach 312–15 optimal portfolios using 313 and hedging to benchmark neutral 328 and setting good benchmark 302–3 some practical considerations concerning 328 performance-based benchmarks, problems with 135–6 Pictet 123 Plaza Accord 3 portfolio insurance 85, 258, 350, 352, 353 prime brokerage 122, 401–3 agreement 144 counterparty banks and 121–3 and currency overlay 196–9 prime brokers and counterparty banks 121–3 proxy hedging 187–8 purchased options 185; see also options purchasing-power parity (PPP) 202, 322 long-run prevalence of 50–1
488
CO2_with_index_2nd_amends.indd 488
23/05/2011 14:10
INDEX
models based on 203–9 absolute 204–5 absolute, problems with 205 application of, examples 207–8 and empirical testing 205–7 and fair-value models 208–9 and law of one price 203–4 relative 205
customised benchmarks and 304–6; see also benchmarks risk-control objectives versus return enhancement of currency-overlay managers 114–15 rollover date 186 Royal Bank of Scotland 121 Russell Investments 124, 127, 146
Q quants 108
S seasonality 242–3 sector-based investment and currency management 298 service providers 78, 117, 141 Sharpe Ratio 59, 250, 256–7, 287, 293, 387 Singapore dollar 296, 386–7 single-bank platforms 408–10 skews and smiles 359–61 Solnik–Sercu currency-hedging model 271–3 currency risk premium in 272–3 South African rand 246, 387 standard deviation 125, 129, 257, 288, 293, 363 Stark Currency Index 126 State Street 55, 112, 123, 146, 215 statistics, a word of caution concerning 39–40 stochastic volatility 356 stock prices and currency fluctuations 36–8 stop-loss and take-profit 229–30 straight-through processing 404
R random walks, currency movements as 23, 42–3 Rating and Investment Information 124 rational expectation model 211 RBC Dexia 123 real effective exchange rate (REER) 208, 322 Record Currency Management 112, 119, 250 redundancy of options 355–8 relative strength index (RSI) 117 reserve accumulation 400 return-enhancement versus risk control objectives of currency overlay managers 114–15 risk appetite, and indexes of risk 388–90 risk constraints 166–72 different restrictions for different types of currency programmes 171–2 leverage limit 168–9 “no net short sale”, “no leverage” and other position limits 166–8 stop-loss and maximum-drawdown limits 170 tolerance range around hedge ratio 170–1 value-at-risk limit 169–70 volatility and tracking-error limit 168 risk premiums 43, 46–7, 272–3 risk separation and risk responsibility 93
T Thai baht 387, 388 Thompson Reuters Matching 397 time-varying risk premiums 49, 224 Tokyo Pension Research Institute 124 trading spot, using option skew as an indicator 368–9 trading styles: active and passive, choice between 156–7
489
CO2_with_index_2nd_amends.indd 489
23/05/2011 14:10
CURRENCY OVERLAY – SECOND EDITION
for active overlay and currency pure alpha 104–9 classification of 106 fundamental-discretionary 108 fundamental-quantitative 106–7 technical-discretionary 107–8 technical-quantitative 108–9 for currency-overlay managers 117–18 transparency and transaction cost 6–7 tree methods 236
W Warburg Dillon Read 211 Watson Wyatt 127 Wealth Preservation Currency Index 297 wings volatility 368 WM/Reuters rates 188–9 World War One 36 World War Two 2
U UBS 44, 121, 207, 211, 215 investment categories of 45 unfunded pure alpha 103–4 unitary hedge ratio 281–2; see also hedge ratios US dollar: continuing dominant currency 11 cycle 333 depreciation period of 79–80 downward trajectory of 4 hegemony of 5 Hong Kong dollar pegged to 92 steady appreciation of 3 world’s most important currency 290 V value-at-risk (VAR) 53, 169–70, 240 variance–covariance matrixes, instability of 289–90 Velocity 397 volatility: actual 72, 360, 362–4, 366, 368 and correlation forecasting 365–6 implied 72, 345, 362–3, 365–6, 368, 385–6 research on behaviour of 385 selling 368 trading schemes based on 366–9 gamma scalping 367–8 trading spot using option skew as an indicator 368–9 wings 368 variance swap 367
490
CO2_with_index_2nd_amends.indd 490
23/05/2011 14:10