reading

Chapter 8: Statistical and Economic Methods for Evaluating Exchange Rate Predictability

Chapter 9: When Are Pooled Panel-Data Regression Forecasts of Exchange Rates More Accurate than the Time-Series Regression Forecasts?

Chapter 16: Active Currency Management Part I: Is There a Premium for Currency Investing (Beta)

Chapter 17: Active Currency Management Part II: Is There Skill or Alpha in Currency Investing?

17.4 Empirical Evidence with the Barclay Currency Traders Index and Individual Fund Managers

23.3 Dislocations of the FX and Cross-Currency Swap Markets under Financial Crises

Chapter 24: FX Options and Volatility Derivatives: An Overview from the Buy-Side Perspective

26.2 Official FX Interventions and Reserve Accumulation: Stylized Facts, Motives, and Effects

No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/permission.

Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.

For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com.

Handbook of exchange rates / [edited by] Jessica James, Ian W. Marsh, Lucio Sarno.

The Foreign Exchange market, as we know, was born in the 1970s. To us in the second decade of the third millennium, mainly accustomed to freely floating and convertible currencies, the events that led to its birth would seem almost incredible. It is huge news when a government, via its central bank, intervenes to shift its currency or keep it stable. But from the end of World War II to the mid-1970s, pegs and interventions were the norm for almost all currencies.

The post war era gave us the Bretton Woods agreement that set the United States as the world's reserve currency and pegged most others to it. The United States was backed by gold, and the stability that this system brought enabled trade flows to enormously grow worldwide. This was deliberate; in the preceding decades, the architects of the system had seen how economic stresses could lead countries to war and were driven by a desire to promote trade and allow the world economy to grow.

However, the Vietnam war brought a growing US trade deficit and a drift of the alignment of the United States and the US dollar with the rest of the world. Pressure grew on the exchange rates and one by one, led by the Japanese yen, the currency became freely convertible and floating. By 1976, the gold standard was no more present and the majority of the world's currencies had the form we know today.

From that not-so-auspicious beginning grew the world's largest market. Today, according to the 2010 BIS report, currencies to the value of over US$4 trillion are traded each working day in global markets. The total value of global equity is a fraction of this amount. In most countries, the traded volume of the currency vastly exceeds the total amount of government debt. Foreign exchange has indeed come of age.

But what is it like, compared to other rates? It is not an interest rate. An interest rate is highly predictable and stable between close to zero, and perhaps exceptionally 15%. Outside these ranges, interest rates historically do not last long; high rates are often followed by spiralling hyperinflations, and negative rates do not occur without anomalous market stresses. It is not an equity, which has a floor at zero and is at least expected to appreciate with inflation.

There is a symmetry to foreign exchange, which is lacking in other market rates. One is “long” or “short” of an equity; in foreign exchange to be long of one currency is to be short of another. It is a rate of exchange and not an ownership of an asset. It has no floor or ceiling, no particular “stable region.” As befits the largest global market, it is also the easiest and cheapest to trade in, with transactions often costing just a single basis point—1% of 1%—in bid-offer spread.

How and why has it grown to such a size? Partly, it is the nature of the beast and the way it is traded. Most foreign exchange transactions are done on a margined basis, with the principal amounts being notional only. Thus, for a simple forward trade of 1 month's duration, where a rate of exchange is locked in for 1 month ahead, no capital is exchanged at the start. At the end, the difference between the agreed rate and the market rate is paid out, scaled by the notional amount of the deal. This is usually at most a few percent of this notional amount. So one can see that it is unnecessary to actually own a sum of money equal to the notional amount—it is more of a scaling factor to the deal. Hence, the notional flows in the market can be very large relative to a smaller capital base.

Another reason for the size of the foreign exchange flows is that speculative trades, which have grown vastly in size and number over the last few years, are usually leveraged. The foreign exchange market is somewhat less volatile than the equity or commodity market, and so to generate comparable profits, profit seeking traders usually apply a multiplier to the notional amount. The last decade has seen the development of foreign exchange as an asset class, where structures suitable for investors are created around foreign exchange transactions. These often utilize large notional amounts to generate returns.

These additions increase an already considerable bulk of foreign exchange trades in the market. One critical activity made possible by the advent of floating exchange rates is hedging. Overseas investments can have their value drastically altered by fluctuations in the foreign exchange rates; hedges allow investors and corporates to immunize their portfolios and cashflows against these variations. Thus the growth of overseas investment has itself driven the volume of foreign exchange transactions higher; the investments, particularly less volatile fixed income instruments, will be significantly affected by foreign exchange movements unless they are hedged.

What are the instruments that make up this flow? The vast majority are spot and forward transactions, but the option market is also deep and liquid. This market came into its own in 1983 when Garman and Kolhagen published the formula for an option on a foreign exchange rate. Additionally, a vast variety of structures may be constructed, which allow investors or corporates to eliminate or transform risks, or gain exposure to specific areas and events.

Why do we need a Handbook of Exchange Rates? In part because it is so important, and in part because it is evolving so rapidly. In the last 5 years, we have seen the growth of foreign exchange as an asset class, which 10 years ago was the exclusive property of large trading floors and a few fund managers. Most large investors will now have a portion of their risk allocated to active strategies, which deliver fundlike returns derived from positioning themselves in the foreign exchange market. Another significant change is the growth of correlation and volatility derivatives. The correlation of rates has become a traded quantity, allowing a remarkably subtle set of risks and rewards to be accessed. The fascinating dynamics of high frequency data are now largely understood via scaling laws, to the extent that co-location—the siting of trading hubs close to exchanges—is becoming an issue; the speed of light has become a limiting factor in foreign exchange trading.

Another fascinating development is strongly connected with the correlation properties of foreign exchange rates. In floating emerging markets, the majority of foreign exchange rates are closely connected with their equity markets for obvious reasons. So high is this correlation that one can replicate emerging market equity indices to good accuracy using only foreign exchange rates—with their greater liquidity and low trading costs. Other interesting correlations are those between the Japanese yen or Australian dollar to the world's equity markets. So strong is the negative (in the case of the Yen) or positive (the Australian dollar) correlation that currency positions are often used to hedge equity drawdowns. So the foreign exchange market is being used to proxy for other rates. Its liquidity and depth make this attractive, though market stresses could bring an end to some of these useful correlations.

The Handbook is designed to span this extensive subject, with experts in the different areas contributing to each section. In planning the book, we drew up a list of the key subjects deserving of a chapter and against each we wrote the name of the key people in that topic. In almost every case, to our delight, those people accepted our invitation to contribute. As a result, the chapters have been written by leading specialists in their fields, often with extensive experience in academia and/or professional practice. Each chapter then benefitted from the feedback of at least one anonymous referee and at least one coeditor, and most authors kindly also acted as referees.

The initial section “Overview” covers structure, regimes, and general underlying behaviors. Part 2 covers models and methods, focussing on the predictability—or lack of it—of foreign exchange rates. In Part 3, we go to the practitioner side of the subject to cover hedging, active management, high frequency trading, and products. Part 4 wraps up with a look at policy and a framework for analyzing currency crises. The chapters are mainly in the form of self-contained surveys, and trace the key developments in a well-defined topic with specific reference to the relevant research frontier. Some contributions also present new empirical findings, especially where competing paradigms are evaluated. So rapid has the evolution of this market been, that it would not be surprising if in a few years there are many new chapters to add! Alternatively, market stresses and political crises can influence foreign exchange rates to an enormous extent. It is easy to forget that our current regime of floating market determined rates is recent, and by no means the norm in a historical context. The foreign exchange market is a fascinating subject, which never stops evolving.

We owe a debt of gratitude to many people for helping us bring this project to a successful conclusion: the contributing authors, for the high quality of their chapters; the many anonymous referees who provided, without exception, valuable feedback to the authors; and Chiara Banti who collated the chapters and worked very hard to copy edit the first draft of the Handbook. Needless to say, without their help, we would never have completed this project.

Saeed Amen, Nomura, London, UK

Naohiko Baba, Goldman Sachs Japan Co. Ltd., Tokyo, Japan

Philippe Bacchetta, Department of Economics, University of Lausanne, Lausanne, Switzerland; CEPR

Craig Burnside, Department of Economics, Duke University, Durham, NC, USA; CEPR; NBER

Gino Cenedese, Bank of England, London, UK

Yin-Wong Cheung, Department of Economics and Finance, City University of Hong Kong, Hong Kong

Menzie Chinn, Robert M. La Follette School of Public Affairs; Department of Economics, University of Wisconsin-Madison, Madison, WI, USA; NBER

James E. Dalton, CitiFX Intelligent Orders, Citigroup, London, UK

Paul De Grauwe, Department of Economics, University of Leuven, Leuven, Belgium

Pasquale Della Corte, Imperial College, London, UK

Alexandre Dupuis, Olsen Ltd., Zurich, Switzerland; Centre for Computational Finance and Economic Agents (CCFEA), University of Essex, UK

Martin D. D. Evans, Department of Economics, Georgetown University, Washington DC, USA; NBER

Robert Flood, University of Notre Dame, Notre Dame, IN, USA

Jeffrey A. Frankel, Kennedy School of Government, Harvard University, Cambridge, MA, USA; NBER

Marcel Fratzscher, European Central Bank, Frankfurt/Main, Germany

Kristjan Kasikov, CitiFX Quantitative Investor Solutions, Citigroup, London, UK

Geoffrey Kendrick, Nomura, London, UK

Michael R. King, Richard Ivey School of Business, University of Western Ontario, Canada

Richard M. Levich, Finance Department, Stern School of Business, New York University, New York, NY, USA

Hanno Lustig, UCLA Anderson School of Management, Los Angeles, CA, USA; NBER

Nancy Marion, Dartmouth College, Hanover, NH, USA

Nelson Mark, Department of Economics, University of Notre Dame, Notre Dame, IN, USA; NBER

Ian W. Marsh, Cass Business School, London, UK

Paul R. Masson, Rotman School of Management, University of Toronto, Toronto, Canada

Christopher J. Neely, Research Department, Federal Reserve Bank of St. Louis, St. Louis, MI, USA

Lawrence H. Officer, University of Illinois at Chicago, Chicago, IL, USA

Richard B. Olsen, Olsen Ltd., Zurich, Switzerland; Centre for Computational Finance and Economic Agents (CCFEA), University of Essex, UK

Carol Osler, Brandeis International Business School, Brandeis University, Waltham, MA, USA

Frank Packer, Bank for International Settlements, Basel, Switzerland

Evgenia Passari, Cass Business School, London, UK

Michael J. Paulus, Hongkong and Shanghai Banking Corporation (HSBC), Hong Kong

Momtchil Pojarliev, Hathersage Capital Management LLC, South Norwalk, CT, USA

Dagfinn Rime, Research Department, Norges Bank, Oslo, Norway; Norwegian University of Science and Technology

Pablo Rovira Kaltwasser, University of Leuven, Leuven, Belgium

Yuji Sakurai, UCLA Anderson School of Management, Los Angeles, CA, USA

Lucio Sarno, Cass Business School, London, UK; CEPR

Aysu Secmen, CitiFX Quantitative Investor Solutions, Citigroup, New York, USA

Thomas Stopler, Goldman Sachs, London, UK

Donggyu Sul, University of Texas, Dallas, TX, USA

Oleg Svirschi, Record Currency Management, London, UK

Ilias Tsiakas, Department of Economics and Finance, University of Guelph, Guelph, Ontario, Canada

Eric van Wincoop, Department of Economics, University of Virginia, VA, USA; NBER

Adrien Verdelhan, Department of Finance, MIT Sloan School of Management, Cambridge, MA, USA; NBER

Paul A. Weller, Tippie College of Business, University of Iowa, Iowa City, IA, USA

Juan Yepez Albornoz, Department of Economics, University of Notre Dame, Notre Dame, IN, USA

1.1 Introduction

It would be hard to overstate the importance of exchange rates for the world economy. They affect output and employment through real exchange rates. They affect inflation through the cost of imports and commodity prices. They affect international capital flows through the risks and returns of different assets. Exchange rates are justifiably a major focus for policymakers, the public, and, of course, the media.

To understand exchange rates, it is essential to know how these prices are determined. This chapter describes the FX market and presents new evidence on recent trends, thereby setting the stage for the rest of the handbook. It presents stylized facts on the market's size and composition. It then describes more closely the motives, incentives, and constraints of the major players. Trading is a search problem, and the constraints and costs related to this search are affected by the structure of the market. Our starting point is that the exchange rates are primarily driven by new information about economic fundamentals. In this light, we review which agents bring information to the market and exactly how their information becomes embedded in the market price.

The chapter describes the momentous changes in trading practices and market structure which have taken place over the recent decades. It concludes by presenting new evidence on some of the most recent technological advances. Twenty-five years ago, most FX trading involved the telephone and all trading involved institutions: individuals were essentially shut out. Trading was opaque, there was a sharp division between interdealer trading and dealer–customer trading, and market concentration among dealers was low.

Today, only the least liquid corners of the FX market can still be described this way. In the liquid markets, telephones are obsolete. New electronic trading platforms have streamlined trade processing and settlement, reduced operational risks, and lowered trading costs. Lower trading costs have enabled the participation of retail traders and the adoption of new strategies such as high frequency trading—a form of computer-automated trading that relies on high execution speeds to make profits from small price movements. Since streaming real-time prices are now available to virtually all participants, these markets are now regarded as transparent. On some of the new platforms, any trader can provide liquidity, so the division between dealers and their most sophisticated customers is, at times, ambiguous. To remain competitive, the major dealing banks have made heavy investments in software and hardware, which has led to sharply higher concentration among market makers.

1.2 Geography and Composition of Currency Trading

Given the pervasive influence of exchange rates, it is no surprise that the dollar value of trading activity in spot and forward FX market dwarfs most other economic measures (BIS, 2010). With daily average turnover estimated at $2.0 trillion, the market is 36 times larger than the combined exports and imports for the world's 35 largest economies, 16 times their combined GDP, and roughly 10 times exchange-traded equity turnover.

FX trading volume has exploded reflecting an electronic revolution that has lowered trading costs, attracted new groups of market participants, and enabled aggressive new trading strategies. Between 1998 and 2010, turnover in the FX market grew by over 250% (BIS, 2010). The associated 8.4% average annual growth rate far exceeds the contemporary 5.5% annual expansion of global real GDP (Table 1.1).

Table 1.1 FX Turnover and Growth: Comparison with Trade, GDP, and Equity Trading Volume

Many aspects of the FX market remain constant despite the electronic revolution. As has been true for decades, the markets remain decentralized with high liquidity and continuous trading (Lyons, 2001; Rime, 2003; Osler, 2009). As ever, the trading day begins when dealers arrive for work in Australia and Asia. Activity then moves to Europe when markets open in Frankfurt, London, and Paris, and finishes late in the afternoon in New York. As always, there is no time during the day when the market formally closes, although there is a brief lull in activity between about 19:00 and 22:00 GMT, when most New York traders have gone home and most Sydney traders are still on their way to work (Figure 1.1). As ever, overall market liquidity is deepest when both London and New York are open, though liquidity for most individual currencies tends to be deepest during their local trading hours.

FX trading remains heavily concentrated in London, which captures over one-third of global trading, and New York, which captures almost one-fifth of trading (Table 1.2). London's traditional dominance in FX grew out of the United Kingdom's worldwide economic dominance in the nineteenth century. It remains secure at the beginning of the twenty-first century because of its geographic location: London's morning session overlaps with Asian trading and its afternoon session overlaps with New York's trading. Trading in the Asia-Pacific region, which in aggregate accounts for about one-quarter of global trading, is dispersed among Tokyo, Hong Kong, Singapore, and Sydney. Latin America, Africa, and the Middle East each account for less than 1% of global turnover.

Table 1.2 Geographical Distribution of Global Foreign Exchange Market Turnover (%)

Despite the continued dominance of London and New York, there have been some subtle shifts in the global distribution of currency trading. The advent of the euro (EUR) brought a decline in the share of European trading outside of London. Meanwhile, rapid economic growth in Asia has supported a surge in trading in the Asian regional centers. Hong Kong and Singapore now vie in importance with traditional European centers such as Switzerland and France.

1.2.1 Which Currencies Are Traded?

Another unchanging aspect of the FX market is the dominance of the US dollar (USD), which is still involved on one side of roughly three-quarters of all spot transactions (Table 1.3).¹ The dollar's dominance reflects the market's practice of trading minor currencies via a major currency (called the vehicle currency). A trade from Mexican pesos (MXP) to Australian dollars (AUD), for example, would typically involve two trades, one from MXP to USD and the second from USD to AUD. This “vehicle” trading through the major crosses concentrates liquidity in a narrower range of currency pairs, reducing overall transaction costs.

The EUR is involved in 46% of trades, in part because it serves as the vehicle currency within the eurozone.² The next most actively traded currencies are the JPY (20%) and the UK pound (GBP, 14%). Together, these four currencies are known as “the majors” (or G4).³

The next tier below the majors comprises the AUD (7.5%), the Swiss franc (CHF, 6.2%), and the Canadian dollar (CAD, 5.2%). A notable recent shift is the rising share of the so-called commodity currencies, specifically the AUD, CAD, NOK, and the New Zealand dollar (NZD). These currencies' combined share rose from 7% in 1998 to 16% in 2010.

The share of emerging-market currencies rose sharply in the 1990s but has been fairly stable around 18% since then. Nonetheless, currencies from the most advanced emerging markets, such as the South Korean won (KRW) and Hong Kong dollar (HKD), have more than doubled their market share since 1998 and now rival the Swedish krona (SEK). Turnover in more recently emerging countries, such as Turkey, Thailand, Brazil, and India, has grown even faster.

The conventions governing the quotation of different currency pairs have also been fairly stable over time. Most exchange rates are expressed as units of a given currency required to purchase 1 USD. The exceptions are the EUR, the GBP, the AUD, and the NZD, which are quoted as the base currency (i.e., EUR/USD = USD per EUR). Most exchange rates are quoted to five significant digits, with the final (or smallest) digit known as a “pip.”⁴

1.2.2 What Instruments Are Traded?

The dominance of spot FX trading is another area of relative stability.⁵ Daily spot turnover in 2010 was $1.5 trillion while turnover in outright forwards was far lower, at $0.5 trillion (BIS, 2010). A number of other currency-related instruments—FX futures, currency options, FX swaps, and currency swaps—swell total daily turnover in FX market beyond $4.0 trillion (Table 1.4).⁶ These assets are traded entirely separately from spot and forward contracts and for entirely different purposes, so they generally have little influence on exchange rates and are not discussed in this chapter.

FX swaps deserve some discussion, however, because of their immense average daily turnover of $1.8 trillion. Like repos in the fixed income markets, FX swaps are primarily used for overnight position management by banks. Collapsing swap volumes following the Lehman Brothers bankruptcy in late 2008 triggered a rapid expansion of central bank swap activity, as authorities tried to stabilize the world banking system (Baba and Packer, 2009; Melvin and Taylor, 2009). In 2010, FX swap trading activity remained below its previous peak.

1.2.3 How Is Trading Regulated?

The vast majority of FX trading is essentially unregulated, in striking contrast to the extensive regulations in most equity and bond markets.⁷ Governments have learned through experience that dealers will simply move elsewhere if they are regulated. In the 1960s, for example, bond dealers moved offshore when the US government attempted to regulate the foreign issuance of USD denominated bonds in the domestic market.

Some well-known regulations in other asset markets are missing in FX market. Their absence is not a problem because of unique features of the FX market. Short-sales restrictions, for example, though severe in most developed equity and bond markets, cannot even be defined in this asset class because the sale of one currency is simply the purchase of another. Other practices that are illegal on most organized exchanges are discouraged in FX by market conventions and best practices. For example, front running of customer orders is widely considered bad practice even though it is not illegal.⁸ Fortunately, the FX market is sufficiently liquid that significant manipulation by any single actor is all but impossible during active trading hours for the major currencies.

Since FX market are subject to minimal regulation, they are also subject to minimal reporting requirements, which explains the scarcity of aggregate data on FX trading. Although equity trading volume is a staple on the evening news on any given day, no one knows how much was traded in FX market—not the regulators, not the monetary authorities, and not even the major FX dealers. The only comprehensive information source is the Triennial Central Bank Survey of FX Market Activity, a joint effort of central banks around the world, coordinated every third year in April by the Bank for International Settlements (BIS).⁹ In the absence of official sources of high frequency data, most research on currency trading relies on proprietary data from banks and brokers.

1.3 Players and Information in FX Market

A key goal of exchange rate economics is to understand currency returns. Exchange rates, like asset prices, more generally, move in response to new information about their fundamental value. Over the past decade, microstructure research has revealed that this “price discovery” process involves different categories of market participants. Each participant's distinct role is determined by (i) whether the agent is a liquidity maker or taker and (ii) the extent to which the agent is informed.

The original FX market participants were traders in goods and services. Currencies came into existence because they solved the problem of the coincidence of wants with respect to goods. Most countries have their own currencies, so international trade in goods requires trade in currencies. The motives for currency exchange have expanded over the centuries to include speculation, hedging, and arbitrage, with the list of key players expanding accordingly. Beyond importers and exporters, the major categories of market participants now include asset managers, dealers, central banks, small individual (retail) traders, and, most recently, high frequency traders.

“Dealers” or “market makers” emerged naturally to fulfill the search function among trading counterparties. Dealers stand ready to trade with anyone needing FX at a moment's notice. To initiate an FX trade, an agent calls a dealer indicating the currency and quantity she/he wishes to trade and asking for the price. The dealer states a price at which she/he is willing to buy (the “bid”) and a price at which she/he is willing to sell (the “ask”). Finally, the customer decides whether to buy, sell, or pass. The dealer is compensated for the burdens of liquidity provision, such as bearing inventory risk and screening agents for credit quality, by a favorable gap between the quoted buy and sell prices, the “bid–ask spread.” Markets of this structure, known as “over-the-counter” (or OTC) markets, have arisen naturally in contexts including municipal and corporate bonds, derivatives, and equities. Although OTC dealers are under no formal obligation to provide liquidity, they tend to be reliable because otherwise their reputation—and potentially their market share—will suffer.

Existing theory indicates that information is brought to the market by liquidity takers rather than market makers. Among liquidity takers in FX, the evidence indicates that information relevant to short horizons comes from financial customers, especially leveraged investors. Information does not come from firms involved in international trade, small individual traders, or governments/central banks (Bjønnes et al., 2005, 2011; Evans and Lyons, 2006; Nolte and Nolte, 2009; Osler and Vandrovych, 2009; King et al., 2010). To identify whether some category of participant is typically informed, it is standard to examine whether their trades anticipate FX returns. If an agent consistently tends to buy (sell) before prices rise (fall) and the subsequent price change tends to be at least partially permanent, researchers infer that the agent was trading based on information about the asset's fundamental value. Academic studies have long indicated that FX dealers are informed. But, until recently, both theory and practice assumed that this information originated entirely with end-customers (Evans and Lyons, 2002). It is now recognized that dealers bring their own independent information to the market (Bjønnes et al., 2011; Moore and Payne, 2011).

1.3.1 Who Needs Liquidity?

Traditionally, the end-customers routinely needing FX liquidity were non-dealer financial institutions on one hand, and corporations and governments on the other. Over the last decade, the set of active FX end-users has been augmented by retail investors and by computer-automated traders known as algorithmic traders.

Until the mid-1980s, non-dealer financial institutions, corporations, and governments each traded roughly equal amounts with their FX dealers. Since then trading by corporate customers and governments has maintained an overall market share of 17% on an average, though this share naturally rises and falls with economic activity. During the recessions of 2001 and 2010, for example, their share of activity fell to 15%. The share of financial trading in total trading, by contrast, rose steadily from 20% in 1998 to over 50% in 2010 (Figure 1.2). This trend partly reflects the rapid growth of trading on retail trading platforms,¹⁰ which reached an estimated $125 to 150 billion per day in 2010, equivalent to 8–10% of global spot turnover (King and Rime, 2010). It also reflects rapid growth in algorithmic trading, especially, high frequency trading. Although data on the extent of algorithmic trading are limited, the survey reported in Section 1.1 suggest that it now accounts for between one-third and one-half of trading in the most liquid currencies.

Financial Institutions

Financial institutions are a diverse category that includes hedge funds and other asset managers, regional and local banks, broker–dealers, and central banks. Relative to corporate customers, financial institutions trade larger amounts and hold FX positions for far longer. Financial institutions tend to be better informed than other end-users, because they have strong incentives to invest in information acquisition. Since financial institutions use currencies primarily as a store of value, they gain or lose according to changes in the currency's value.

Among financial institutions, leveraged institutional investors—meaning mostly hedge funds and their close cousins the commodity trading advisors (CTAs)—appear to be best informed. This finding seems logical since leveraged institutional investors face particularly intense incentives to acquire information. Hedge-fund managers are often paid 2% of underlying assets as a flat annual fee and 20% of investment returns. Leveraged currency funds, which grew dramatically during the late 1990s, are known to favor well-defined speculative strategies focused on four factors: fundamentals, interest differentials (i.e., the carry trade), momentum, and volatility.

Unleveraged asset managers (“real-money investors”) include mutual funds, pension funds, endowments, and insurance firms. Perhaps surprisingly, such funds often pay little attention to the exchange rate component of returns when choosing asset allocations (Taylor and Farstrup, 2006). Instead, they concentrate on maximizing expected returns to foreign assets measured in the asset's home currency. This approach may be rational given ample evidence that major exchange rates are well approximated by a random walk. Some real-money investors outsource the management of FX exposures to currency overlay managers, who focus on risk reduction, return maximization, or some combination of the two.

FX hedging has become more important among portfolio managers since the financial crisis (Melvin and Prins, 2010). Market participants report that it is common to adopt a 50% hedge ratio, with the hedge reset periodically (e.g., once a month). A 50% ratio minimizes “embarrassment risk,” meaning the risk that a firm incurs either an absolute loss (when the rate moves adversely on an unhedged position) or an opportunity loss (when the rate moves favorably on a fully hedged position).

Private financial institutions dominate financial trading on a day-to-day basis, but central banks are noteworthy participants nonetheless. When these public sector institutions intervene to influence exchange rates, their trades are considered informed. Major dealing banks ensure that they know of such trades by cultivating their relationships with central banks. For example, dealers may share market intelligence on a daily basis with these valued clients. Central banks also trade FX as part of the regular procurement process for military and other government functions. Such trades are not considered informative, and central banks often ensure that such trades are not confused with intervention by announcing them in advance.¹¹

Corporate Customers

Corporate customers use FX market to support treasury operations associated with their core business activities such as mining, shipping, and manufacturing. As such, corporations primarily use foreign currencies as a medium of exchange, trade relatively small amounts, and hold these positions only briefly. Most corporate customers choose not to engage in speculative FX trading; indeed some firms explicitly prohibit it. Given their institutional goals, this restriction seems logical. FX forecasting is not among a corporation's “core competencies,” so cultivating in-house speculative expertise can be ill-advised (Goodhart, 1988). Further, creating a trading operation is expensive. Not only it is costly to hire currency analysts and traders but also it is expensive to hire the extra staff required to protect against “rogue trader risk,” meaning the risk that a single trader brings down the firm (Osler, 2009).¹² Even corporate firms that hedge their foreign cash flows pay little attention to future exchange rate movements. A survey by Bodnar et al. (1998) finds that among corporations that hedge their exposures—as most do—they typically choose hedge ratios between 40% and 50% and favor maturities below 6 months. They also review their currency hedge ratios at most a few times each year. Since corporate customers generally choose not to engage in speculative trading, it is not surprising that their trades do not anticipate short-term returns and are, therefore, not considered informative.

Corporations typically only use the FX market for one side of each exposure. A US multinational needing EUR to pay taxes in Germany, for example, sells its USD to buy the EUR in the FX market but then delivers the currency directly to the German government, bypassing the FX market entirely. Similarly, a Japanese exporter of manufactured goods to the United States receives USD from the American importer and then sells those USD in the FX market.

Retail Investors

Historically, few private individuals have had sufficient net worth to qualify for a credit line at a FX dealing bank. This barrier to entry effectively made the FX market an entirely wholesale market. Trading by small investors was also discouraged by the relatively high bid–ask spreads on small trades, meaning those below $1 million. Retail investors gained access to FX market around the year 2000 with the arrival of internet-based trading platforms tailored to their unique needs, so-called “retail aggregators” (RAs, described below).

Retail investors primarily trade FX spot in the major currency pairs, although the number of emerging-market currencies offered is growing. These individuals or small institutions tend to focus on just one or two currencies and to hold positions for very short time horizons, typically under a day. According to a recent survey (CitiFX Pro, 2010), these traders find the FX market attractive in part because of its low correlation with other markets, its high liquidity, and its 24-h market.

Retail traders should have strong incentives to be informed, since they trade for speculative purposes and employ substantial leverage. The evidence indicates, however, that retail trades are not informed. Their trades do not generally anticipate exchange rate returns (Nolte and Nolte, 2009) and the retail traders themselves are generally unprofitable (Heimer and Simon, 2011). In 2011, Oanda.com claimed that 48% of their retail customers were profitable. A systematic lack of trading acumen also appears to characterize retail traders in equities (Barber and Odean, 2000, 2002; Linnainmaa, 2010). Well-documented forces that might drive traders to stay active even when losing money include wishful thinking and overconfidence (Oberlechner and Osler, 2012).

Algorithmic and High Frequency Traders

Algorithmic trading is a form of electronic trading where a computer algorithm (or program) determines an order-submission strategy and executes trades without human intervention (Chaboud et al., 2009). Human involvement is limited to designing the algorithm (or algo), monitoring it, and occasionally adjusting the trading parameters. Some algorithms simply automate existing strategies—for example, they break up large trades to minimize transaction costs—while others take advantage of superior execution speeds such as high frequency trading.