SCHOOL OF FINANCE AND ECONOMICS

SCHOOL OF FINANCE AND ECONOMICS UTS:BUSINESS WORKING PAPER NO. 107 SEPTEMBER, 2000 Trading in the Australian Foreign Exchange Market Tiffany Hutcheson ISSN: 1036-7373 http://www.business.uts.edu.au/finance/

TRADING IN THE AUSTRALIAN FOREIGN EXCHANGE MARKET Tiffany Hutcheson School of Finance and Economics University of Technology Broadway, NSW 2007 Australia Abstract The market for foreign exchange is the most heavily traded financial market. Theoretically exchange rate movements are determined by economic fundamentals, such as inflation and interest rates, which influence the supply and demand for currencies. However, empirical models of exchange rate determination based on these fundamentals have not been very successful in predicting exchange rate movements, especially over the short run. Furthermore, market practitioners argue that they have successfully developed profitable trading strategies, which do not rely on an analysis of economic fundamentals. A reason for the poor performance of econometric models and trading strategies based on fundamental analysis could be the attitudes and behaviour of practitioners trading in the foreign exchange market. In order to obtain information about trading behaviour in the Australian foreign exchange market dealers authorised to trade in this market were approached in a survey for their views on market behaviour and trading strategies. This paper provides a detailed analysis of the responses to the survey questions. I would like to thank Warren Hogan, Peter Pontikis and Andrew Spencer for their comments and suggestions. Naturally any errors and omissions are my own. 2

1. Introduction The market for foreign exchange is the most heavily traded of all financial markets. Trading is undertaken by individuals and organisations in support of management of assets and liabilities in different currencies and trade of goods and services between national economies or those who profit from buying and selling currencies at different exchange rates. The average daily turnover in this market is significant with trading taking place at any time of the day across several time zones. In April 1998 the Bank of International Settlements estimated that the global average daily foreign exchange turnover was US$1,490 billion dollars (see Bank of International Settlements (1999)) 1. In Australia the turnover in all currencies was US$47 billion dollars per day with US$27 billion of that being the turnover in Australian dollars (see Reserve Bank of Australia (1998a)). Unfortunately, exchange rates have proven to be one of the most difficult financial prices to estimate. Theoretically a currency's value is determined by the economic fundamentals, such as inflation and interest rates, that influence its supply and demand. If this were so then it would be unprofitable for traders to adopt trading strategies that were not based on these fundamentals. However, empirical models of exchange rate determination have not been very successful in predicting exchange rate movements, especially over the short run (see Taylor (1995)). Furthermore, market practitioners argue that they have successfully developed profitable trading strategies, which do not rely on an analysis of economic fundamentals. A reason for the poor performance of econometric models and trading strategies based on fundamental analysis could be the attitudes and behaviour of practitioners trading in the foreign exchange market (see Krugman (1989)). For example, some practitioners may trade tactically in a way that forces an exchange rate to move away from its fundamental value so that they can create a currency position that becomes profitable once the exchange rate begins to move back towards its original value. In order to obtain information about the behaviour of practitioners in the Australian foreign exchange market dealers associated with banks and others authorised to 1 This turnover is for spot and forward foreign exchange transactions and foreign exchange swaps. The data has been adjusted to account for both local and cross-border accounting. 3

participate in foreign exchange market were approached in a survey for their views on market behaviour and trading strategies. The information required from the dealers was more qualitative than quantitative so there was no need to collect numerical data on their trading behavior. However, the dealer s responses to the survey questions can be used to develop a framework for econometric models on market behavior. Dealers are an ideal group to survey as the majority of trading in the foreign exchange market occurs through dealers. They need to keep up to date with the events that will have an impact on exchange rates and be able to react quickly to changing information and market sentiment in order to be profitable in their trading. As a result their responses to these survey questions will be very informative. The paper is organised as follows. The next part discusses the preparation of the survey and the collection of the survey data. The responses to the questions on background information are then analysed in part three. The fourth part analyses the responses to the questions about market behavior. The contributions from the survey are stated in the conclusion. 2. The Survey Data Surveys were sent to the treasury departments of 56 institutions authorised by the Reserve Bank of Australia to deal in foreign exchange as at 12 th July 1999 2. 59 copies of the surveys were mailed out as three of the major banks, the Commonwealth Bank of Australia, the National Australia Bank Ltd and the Westpac Banking Corporation had active trading desks in both their Sydney and Melbourne offices. Of the 59 surveys, 39 were completed and returned generating a response rate of 66.1%. This is a very high survey response rate as generally response rates to surveys tend to be about 5 to 10 percent with response rates greater than 30 percent very rare (see Alreck and Settle (1985)). The questionnaire is similar to the ones used in surveys undertaken by Cheung and Wong (1999a, 1999b) on dealers located in Hong Kong, Tokyo, and Singapore and Cheung and Chinn (1999a, 1999b) on dealers located in the United States. A summary 2 At the time of the survey there were 60 authorised dealers. However, in some situations two authorities were being used by the same group so only one survey was sent to the group. The Bank of Queensland and Barclays Bank PLC were not included because they were not trading. 4

of the survey questions and responses is provided in the Appendix. The questions were mainly close-ended requiring a box to be ticked as a response. However, some of the questions allowed respondents to specify a response if they felt the existing choice of responses were inadequate. Rarely did respondents feel the need to elucidate on their own response. The use of surveys to collect information is often criticised, as it is difficult to know whether or not respondents will provide honest responses to survey questions. If respondents incorrectly respond to the questions the survey results will not be valid and unbiased. However, this survey should be relatively bias free as the questions chosen do not cover areas which respondents would feel uncomfortable providing answers on. In order to ensure the survey only included questions that dealers would readily respond to, several experienced Australian dealers were asked to comment on the suitability of the questions to be included. Respondents were also encouraged to provide honest responses, as they were able to complete the survey in their own time, at a location of their choice and with the knowledge that no record was being maintained of how they responded to the survey questions. 1. Background Information The first group of survey questions gathered information on the environment that the dealers operated under and the factors they felt influenced how they traded. Panel 1.1 indicates that most of the respondents hold fairly senior executive positions with 38.5% being treasurers/managers and 53.8% being chief/senior dealers. Consequently, the majority of survey responses should reflect their high degree of experience and knowledge of behavior in foreign exchange markets. Although this survey is undertaken on dealers authorised to trade in the Australian foreign exchange market, only 33.3% of the respondents have their foreign exchange headquarters located in Australia (see panel 1.2). The next popular location is Japan with a 20.5% share followed by Europe with 15.4% and the United Kingdom and Asia (excluding Japan) both having a share of 10.3%. Panel 1.3 indicates that most of the financial institutions employing the respondents had daily turnover, as measured in Australian dollars, of less than 100 million. 28.2% of respondents indicated a 5

turnover between 100 million and 500 million, 17.9% between 500 million and 1 billion, and 7.7% between 1 billion and 5 billion. 3.1 Dealers and brokers Dealers can trade directly with each other or through foreign exchange brokers. Over the past five years there has been a slight increase in the proportion of trading taking place through brokers. Panel 1.4 shows that five years ago on average 46.3% of trading was undertaken directly between dealers and 53.7% through brokers while today an average of 42.4% occurs between dealers and 57.6% through broking. Trading between dealers, often referred to as the inter-bank market, involves dealers calling each other to ask for exchange rate quotes. The quoting dealer will provide a two-way quote consisting of a buying, bid, and selling, ask, quote. If the calling dealer agrees to buy or sell at the quoted exchange rate a trade will occur. The quoting dealer provides this service on the assumption that they will receive the same service in return. Unfortunately, the quoting dealer may be forced to trade at a time when they do not want to trade. Furthermore, the calling dealer might not always be trading at the best quote currently available in the market. Foreign exchange brokers overcome these uncertainties by providing the service of bringing buyers and sellers together in return for a commission or fee. Brokers will search the market for the best bid and ask quotes for their clients, who often are dealers. Traditionally this search has involved the broker contacting several dealers via the phone to ask them to anonymously provide a bid or offer quote for a potential trade. They will then provide the client with the best quote that they have received and if the client accepts the quote a deal will take place. The broker only reveals the identity of the buyer and seller to each other after the deal has been made. Over the past five years technology has had a big impact on how broking services are provided. Instead of using voice broking, where brokers communicate with dealers over the phone, broking can now also be done electronically through service providers such as Reuters and EBS. While five years ago most broking was done traditionally, today a significant amount of broking is done electronically (see panel 1.4). The increased transparency of trades under electronic broking can be seen as one of the major reasons behind this shift. This transparency occurs because electronic broking 6

services provide dealers with screens that show the best bid and ask quotes for several currency pairs. Once again quotes are anonymously contributed but this time dealers record their quote on the screen and press a button. The best quotes will be displayed on each dealer's screens and they can accept a quote by simply pressing a button and a deal ticket will automatically be printed. As all dealers using the service will have equal access to the same information on the best quotes, most trading will take place at quotes close to or the same as those on the electronic broking service. The increased trading transparency makes it difficult for dealers to trade if they provide quotes that differ from those on the electronic broking service. This is because calling dealers would not be prepared to trade at quotes less profitable for them than the quotes from electronic broking. Consequently, quoting dealers would not be able to generate trade in a currency if their bid quote was lower or their ask quote was higher than the current quotes from electronic broking. Furthermore, electronic broking is quicker than voice broking in obtaining the best quotes for their clients so voice brokers are becoming less competitive. As a result you would expect that in the future most broking would be done electronically. 3.2 Dealers and Clients Dealers trade with other dealers in order to generate profits from exchange rate movements or arbitrage opportunities (see Carew and Slayter (1994)). They also undertake trades for clients who want to buy and sell currencies. In these trades dealers may be using their own currency holdings or obtaining the required currencies by trading with other dealers. Whatever their original sources, dealers may enter into offsetting trades with other dealers to maintain desired currency exposures. As a result you might expect the majority of trading in major currencies to occur between dealers. However, the proportion of transactions that a dealer undertakes with other dealers and customers has been fairly equally divided for the past five years (see panel 1.5). This could be due to dealers being able to make substantial profits from satisfying orders from large corporate clients (see Carew and Slayter (1994)). Dealers may find that they can achieve higher returns from servicing these client orders than they can from trading in the highly competitive inter-bank market where profits are not always realised. 7

3.3 Transaction Size Transactions in the foreign exchange market are for large amounts of currency. Panel 1.6 shows that Australian dealers mainly provide quotes on amounts not exceeding ten million Australian dollars 3. 41% of the trades are for less than five million, 30.8% are for five million and 23.1% for ten million. Only 5.1% of the trades are for amounts greater than twenty million dollars. Quotes with dealers in other time zones are still mainly on amounts below five million dollars (see panel 1.7). However, a slightly larger proportion of quotes will be given on trades that exceed twenty million dollars. This could be because trading between Australian dealers tends to be more frequent than trading between Australian dealers and dealers in different time zones so it can support smaller volumes. 3.4 Trading in spot market The factors that influence how dealers trade in the spot foreign exchange market vary between dealers. Panel 1.8 reports that the main factor both today as well as five years ago has been meeting customer orders. In fact meeting customer orders has become even more significant over the last five years with 35.5% of respondents claiming it was a major factor 5 years ago and 45.6% claiming it as a major factor today. The second most popular factor influencing trading over the last five years has been an even mixture of two different approaches to predicting exchange rate movements, technical analysis and fundamental analysis. The use of either technical analysis or fundamental analysis by itself has not had a major influence on trading today or five years ago. Jobbing where dealers profit from differences in dealer quotes had a significant influence on trading five years ago, 22.6%, but its popularity has rapidly decreased to 10.5% today. The preference for using an even mixture of fundamental analysis and technical analysis rather than only one by itself can be easily explained. Fundamental analysis is regarded to be better at explaining long-term exchange rate movements while technical analysis is better at explaining short-term movements. Fundamentalists argue that the value of a currency should reflect the economic fundamentals, such as 3 Intra-day and overnight limits are placed on an individual dealer's currency position. 8

relative inflation rates, interest rates, growth levels and national income, of its country. If the value of one currency against another currency does not reflect these fundamentals, a profitable trading opportunity will exist. Dealers will profit by buying undervalued currencies and selling overvalued currencies. Under floating exchange regimes, if there is no central bank intervention, this arbitrage trading will see the value of a currency move quickly towards its true value (see Neely (1997)). Consequently, exchange rate movements should be random and only occur when new information becomes available. However, most supporters of fundamental analysis confess that economic fundamentals cannot always be used to explain daily fluctuations in exchange rates (see Singleton (1987)). In fact exchange rate movements are not always random but often demonstrate significant trends (see Arnott and Pham (1993)). While this could be due to the market not adjusting automatically to changing economic fundamentals, another reason is that short-term market sentiment has not always been justified by economic fundamentals. For example, dealers could be reacting to recent rumours about events that have not yet occurred or cannot be supported. Technical analysis uses past price movements to estimate trends in future price movements. Under technical analysis dealers develop trading rules based on exchange rate movements, such as buying when exchange rates rise past a ceiling barrier and selling when exchange rates fall below a floor barrier. Consequently, dealers are trading with the trends and identifying signals that indicate when trend lines will break. Taylor and Allen s (1992) survey on trading strategies adopted by London foreign exchange dealers support the popularity of technical analysis in foreign exchange markets. However, dealers in that survey indicated that while technical analysis mainly determined their short-term, intra-day to one week, forecasting fundamental analysis became significantly more important as the time horizon increased. The declining popularity of jobbing can be attributed to the growing difficulty dealers are encountering when attempting to generate acceptable profit levels. Under jobbing dealers profit by buying at low quotes and within a small time frame turning around and selling at higher quotes. However, foreign exchange markets have become more 9

efficient so that arbitrage opportunities like this are occurring less frequently and are of a smaller size than they were five years ago (see Carew (1998)). Technological advancements, such as improved electronic telecommunications and electronic broking, can be seen as a factor partly responsible for this. Today dealers can easily communicate with each other, either by phone or email, regardless of what time zone they are in. They also have ready access via electronic information providers to the current news on domestic and international financial markets and events. Consequently, you would expect all dealers to be trading at very similar quotes, making it difficult for arbitrage trading to take place. Furthermore, as electronic broking provides dealers with equal access to the same currency quotes, the probability that dealers will provide quotes that differ from those provided by the electronic broker is greatly reduced. 2. On the FX Market The questions on market behavior cover the respondents' views on a range of factors that have been identified by past surveys and empirical studies as having an impact on exchange rates. 4.1 Bid-ask spread A bid-ask spread is the difference between a dealer's bid and ask quote. A calling dealer will buy at the quoting dealer s ask quote and sell at their bid quote. Panel 2.1 lists the size of the normal bid-ask spread used by dealers on inter-bank trades for major currency pairs. The majority of respondents claim that the spread size for all the currency pairs is between three to five basis points 4. A minority set the spread for currency pairs other than the Australian dollar/united States dollar and New Zealand dollar/united States dollar at between seven and ten basis points. In panel 2.2 the respondents indicate that the main reason for quoting the normal market spread when trading is to maintain a good relationship with other dealers. Dealers consider that it is more equitable if they provide quotes to each other using the same size spread. This should ensure that they are able to maintain reciprocal trading relationships with each 4 The last decimal place in a foreign exchange quote is referred to as a point. The term basis points is often used when referring to movements in exchange rates and differences between quotes. 10

other. However, dealers do not always quote the normal spread size. To attract sellers they will increase their bid quote and to attract buyers they will decrease the size of their ask quote. The reverse will occur when they want to make their quotes less attractive. The size of their spread will only change if they adjust either their bid or ask quote or adjust both quotes by different amounts. The spread size will remain unchanged if both quotes are adjusted by the same amount. Panel 2.3 shows that most respondents tend to quote a spread size that is similar to or not much larger than the normal spread when trading. Only about 12.1% of respondents indicate that more than 20% of their quotes have a bid-ask spread larger than market practice. However, there appears to be a greater tendency for the respondents to quote a smaller spread than normal with about 29% of respondents indicating that more than 20% of their quotes were smaller than market practice. The popularity of quoting spreads smaller than normal can be explained by dealers being able to attract more trades when they buy currencies at higher than normal prices and sell currencies at lower than normal prices (see Anthony (1997)). While smaller spreads will generate less profit per trade, the increased trade volume attracted by small spreads must be increasing a dealer s overall profits in order for them to continue quoting smaller than normal spreads. The profit potential for dealers quoting large spreads is fairly small as it is difficult for dealers to attract trades if they are only prepared to buy currencies at lower than normal prices and sell currencies at higher than normal prices. In fact if dealers consistently quote large spreads they will become unpopular and eventually find it difficult to generate any trades. So any profitability from quoting large spreads will only be short-term. The respondents indicate in panel 2.4 that customer type and market conditions have slightly more of an influence on spread size than market practice. With regards to customer type, dealers are likely to quote smaller spreads to customers that they do a lot of business with in order to retain these customers. If dealers did not want to do business with a customer they may choose to quote a larger than normal spread in order to deter trade. They could also quote large spreads to dealers who regularly quote them unfavorable spreads as a form of payback. However, dealers would rarely quote larger than normal spreads as it will damage their reputation. This is supported 11

by panel 2.5 were only a small proportion of respondents claimed they quoted spreads which differed from normal to customers they did not want to do business with. Market conditions will influence spread size as they affect a dealer's decision on whether or not to trade. Studies have shown that the volatility, inventory risk and liquidity are market characteristics that have a significant impact on spread size (see Huang and Masulis (1999)). In terms of volatility, dealers will tend to increase their spread size when exchange rates become more volatile as the direction and size of exchange rate movements will become less certain (see Singleton (1987)). By trading with a larger spread dealers can reduce the risk that they will unexpectedly be forced to trade and build up large currency positions (see Bollerslev and Melvin (1993)). It is important for dealers to control their inventory levels in volatile currencies as unfavorable exchange rate movements can prove to be costly. On the other hand if market liquidity increases dealers will need to become more aggressive in their pricing and decrease spread size otherwise they will find it difficult to compete with other dealers (see Huang and Masulis (1999)). The respondents were asked to select from a range of reasons the three (or fewer) most important reasons for quoting a spread size different from market practice. Panel 2.5 indicates that the existence of a thin market or a hectic market was the most popular reason. Customer type, increased market volatility and timing of market news announcements also received a fair amount of support. Dealers holding positions against the market trend or who were experiencing increasing costs by maintaining their position were not very popular reasons. This would be expected, as dealers normally do not want other dealers to know about their trading position. If they start to quote spreads different from market practice during normal market conditions they will essentially be revealing information about their position to other dealers. Other dealers could then proceed to trade in a manner that adversely affects the quoting dealer s position. In thin and hectic markets a higher than normal degree of uncertainty would exist. So it may be preferable for dealers to quote larger spreads than normal. In thinly traded markets only a very small number of trades will occur each day if at all. Consequently, dealers will be uncertain about whether or not a small or large 12

exchange rate movement will occur when a trade eventually takes place. The level of trading in hectic markets is fairly intense with a large number of trades taking place each day. When trading is hectic it will be difficult for dealers to know exactly what is happening to exchange rates as so many trades will be taking place at the same time and dealers will be constantly changing their quotes. Quoting large spreads in thin or hectic markets will allow dealers to reduce the probability of experiencing significant losses following large exchange rate movements. Market uncertainty can also increase prior to news announcements being made. If dealers consider that the available forecasts are fairly reliable they will feel confident when pricing the forecasts into current exchange rates. This could see them decrease their spread size in order to attract trades. However, dealers may not want to trade if they consider the forecasts to be fairly unreliable so will quote a larger spread size. Once the announcement is made dealers are likely to start quoting normal spreads. This is because if the news is different from what was expected dealers will either by mainly buying or selling so will be providing higher bid quotes or lower ask quotes respectively (see Bollerslev and Marvin (1994)). Consequently, their bid and ask quotes will both move in the same direction without the spread size changing. 4.2 Major Participants The average daily turnover in the Australian foreign exchange market has increased over the last couple of years (see RBA (1999a)) 5. At the same time the number of authorised dealers in Australia has decreased from 76 in 1995 to 60 at the time of the survey in July 1999. This decline can be attributed to financial institutions merging, closing or undertaking their foreign exchange activities from non-australian locations. In this type of environment it may be possible for several dealers to dominate the trading in some or all currency pairings. Particularly as the Reserve Bank of Australia reports that 80 per cent of market turnover is generated by the ten largest dealers (see Reserve Bank of Australia (1999a)). However, apart from the currency pairing of the New Zealand dollar/united States dollar, the respondents did not overwhelmingly support trading dominance by major players (see panel 2.6). This 5 The higher turnover could be partly due to speculators using the Australian foreign exchange market as a proxy for markets subject to capital flow crises, such as those located in Asia (see RBA (1999a)). 13

finding is especially relevant for the Australian dollar/united States dollar currency pairing as it has the highest turnover in the Australian market followed a distant second by the United States dollar/japanese yen (see Reserve Bank of Australia (1999a)) 6. The lack of dominance by one or a few major players trading the Australian dollar/united States dollar ensures that trading among dealers should be competitive. The respondent's support for several dealers dominating trading in the New Zealand dollar/united States dollar currency pairing in the Australian foreign exchange market can be easily explained by the close geographical proximity of New Zealand to Australia. The majority of trading in the New Zealand dollar/united States dollar in the Australasian region actually takes place in New Zealand through New Zealand institutions (see Bank of International Settlements (1999)) 7. Only a small amount of trading in the New Zealand dollar/united States dollar occurs in the Australian market. The trading that does occur tends to be undertaken by the few Australian dealers who act for clients undertaking business in New Zealand. In panel 2.7 the respondents indicate that the larger players in the Australian foreign exchange market have a competitive advantage over other dealers due to their ability to attract a large customer base and get better access to information. This is consistent with the larger players being of sufficient enough size to be able to finance the resources required to support a large customer base and a wide range of information sources. Furthermore, it can be difficult for dealers to attract customers away from the larger players who have been able to establish a strong customer base in a particular currency. The ability for large players to employ experienced traders and trade large volumes of currency were also moderately supported by the respondents as sources of competitive advantage. One respondent commented that the large players were able to trade large currency volumes, as their dealers tended to trade under large credit limits. These are limits imposed on the volume of trade that organisations allow their dealers to undertake with a particular counterparty. The credit limit will be more restrictive the higher the credit risk of the counterparty. Larger players tend to impose more 6 In the late 1990s the United States dollar/japanese yen replaced the United States dollar/german Deutsche Mark currency pairing as having the second largest turnover due to introduction of the Euro. 7 As the New Zealand domestic market is very small, even trading in this market tends to be dominated by a small number of New Zealand dealers. 14

restrictive credit limits on dealers from smaller financial institutions thus restricting the trade volume that these dealers can undertake. 4.3 Economic Announcements Under a floating exchange rate regime, exchange rates reflect the value placed on the currencies by market participants. When determining the value of a currency these participants will take into consideration the existing and expected economic fundamentals of the currency s country. So when government announcements of the actual fundamentals differ from market expectations, dealers will trade in a way which causes exchange rates to change. If dealers do not react before or at the same time as other dealers any profitable opportunity that exists will be traded away. Several studies have been undertaken on the response to news announcements in foreign exchange markets, with studies using intra-day data concluding that new economic announcements take only a couple of minutes to have an effect (see Almeida et al. (1998) and Kim (1998)). Panel 2.8 reveals that dealers tend to react within less than a minute when announcements from the major developed countries on domestic inflation, unemployment, trade deficit, current account, interest rates, retail sales and gross domestic product differ from what was expected. One respondent stated that the Tankan, a survey of industrial trends, announcement made by the Bank of Japan was another announcement which the market for the United States dollar/japanese yen reacted to quite quickly. Several respondents claimed that in Australia the Australian dollar/united States dollar spot market reacted instantaneously to announcements by the Australian government whilst the reaction time in other spot market currency pairings was more delayed. The market also did not react as quickly to announcements made by countries other than Australia. This may be partly because the overseas announcements, such as announcements from the United States of America, are made while Australian financial markets, other than the market for foreign exchange, are closed. Consequently, foreign exchange dealers may not fully adjust their trading position until they see the response to the announcements from other financial markets when they open the following day (see Kim (1998)). For example, dealers 15

may wait to see how the Reserve Bank of Australia responds to changes in overseas interest rates. Respondents are asked to select the economic announcement from Australia and the United States of America that they consider has the biggest impact on the Australian foreign exchange market. Panel 2.9 shows that five years ago the respondents regarded interest rate announcements made by the Reserve Bank of Australia as having the biggest impact followed by announcements on the current account, inflation and then trade balance. Today interest rate announcements still have the biggest impact but inflation and unemployment and then the current account and trade deficit now follow it. The strong support for interest rates announcements is expected as the direction and magnitude of international capital flows is affected by changes to relative interest rates between countries. In the past a close link has also existed between the current account and trade balance and the Australian dollar, as the majority of Australian exports are commodities priced in foreign currencies (see Gruen and Kortian (1996)). Consequently, changes to world commodity prices will affect export revenue and exchange rates. A reason for dealers providing greater support for inflation announcements today than they did five years ago could be their growing concern about the sensitivity of international capital flows. The sudden capital reversals experienced by several Asia countries in the late 1990s made dealers very aware of this sensitivity. As the objective of Australian monetary policy is to keep inflation between 2 and 3 per cent, rising inflation will see the Reserve Bank place upward pressure on interest rates (see Reserve Bank of Australia (1996)). Consequently, dealers expect international capital flows will be fairly responsive to announcements of higher inflation. Of the announcements covering economic conditions in the United States, the announcements that have the biggest impact on the Australian foreign exchange market today as well as five years ago are those for interest rates and employment. Current account and trade deficit announcements from the United States have very little impact, as they are not seen to strong indicators of economic and financial conditions in the United States. 16

4.4 Fundamental value Theoretically, if all currently available information on economic fundamentals is correctly priced into exchange rates any movement in exchange rates will only occur when new information becomes available. Consequently, the size and direction of these movements should be random. Unfortunately, the value of the Australian dollar has not always moved randomly against other currencies, particularly intra-daily and daily. In fact there have been times when it has appeared as though dealers have simply disregarded economic fundamentals and are merely overreacting to news and rumours (Shleifer and Summers (1990)). As shown in panel 2.10 the majority of respondents believe that intra-day exchange rate movements do not reflect changes in the fundamental value of an exchange rate. They feel that changing fundamental values are reflected a lot more in the movements that occur within a period of six months or greater. These responses reinforce the belief that fundamental analysis is more appropriate for forecasting exchange rates over a medium to long-term horizon (see Taylor (1995)). The respondents regard excessive speculation and manipulation by hedge funds to be the main factors preventing exchange rates from reflecting their fundamental value (see panel 12.11). Excessive intervention by central banks was the next most heavily supported factor while manipulation by major trading banks and slowness of dealers to respond did not receive as much support. 4.5 Speculation There has been an ongoing debate about whether or not speculation is a stabilising force on exchange rates. Speculators take positions in the foreign exchange market to profit from exchange rate movements. They will buy a currency and so build up a long position if they expect its value to increase and sell a currency to build up a short position if they expect its value to decrease. When the currency's value moves in the expected direction the speculator reverses their position and profits. As speculators will be reversing long positions when exchange rates are increasing and reversing short positions when exchange rates are decreasing, you would expect their trading to offset large upward and downward pressures on the value of a currency. However, as shown in panel 2.12, survey respondents do not unanimously support speculation as a 17

stabilising force with 55.6% indicating that speculation mainly moves exchange rates towards their fundamental value and 44.4% indicate that it moves them away. The lack of support for speculation as a market stabiliser could be due to the occurrence of several speculative episodes since currencies began to be floated in the 1970s. During these episodes excessive speculation has been blamed for driving exchange rates away from their true values as exchange rates increased or decreased by more than was supported by market fundamentals (see Krugman (1989)). In other words it was the speculative trading that was causing the exchange rate to move rather than changing market fundamentals. Consequently, when speculators eventually reversed their position it was some time before exchange rates moved back towards their true value. One reason for the inability of exchange rates to automatically revert back to their true value following excessive speculation could be the phenomenon known as herding behavior (see Banerjee (1992)). Herding occurs when dealers who do not normally speculate start to imitate the trading behavior of speculators without regard to whether or not their behavior is supported by economic fundamentals. It would take these dealers longer to reverse their positions than the large speculators and so exchange rates would continue to be driven away from their true value. Excessive speculation has been one of the approaches to trading adopted by hedge funds. This is not entirely unexpected as hedge funds are typically made up of a group of wealthy investors who employ a manager to achieve extraordinary returns on their funds. That is a return greater than a market's benchmark return. As managers are generously rewarded for their achievements, they adopt aggressive trading strategies (Chicago Mercantile Exchange (1995)). In foreign exchange markets hedge funds have been known to gradually build up very large positions in particular currencies and then reverse these positions rapidly to earn large profits from the excessive exchange rate movements generated. By building up their positions gradually their impact on exchange rates is only really felt when they reverse their positions. These strategies have been used on overvalued currencies subject to a fixed exchange rate regime as well as in markets for floating currencies that are considered by central banks to be operating normally (see Rankin (1999)). 18

The survey respondents would have recently experienced the destabilising impact of hedge funds on the Australian dollar in mid 1998. Between December 1997 and March 1998 hedge funds were said to have built up large short positions in the Australian dollar using borrowed funds (see Rankin (1999)). As the Australian dollar was already experiencing a downward trend caused by the impact of the Asian crisis on Australia's trade and capital flows, the impact of hedge fund selling on exchange rate movements went largely unnoticed. In fact the Reserve Bank of Australia considered market trading and value of the Australian dollar to be compatible with market fundamentals. However, as the Australian dollar approached US60 cents, hedge funds began to sell more aggressively and generated uncertainty among other traders about what exchange rate the Australian dollar should trade at. In fact parties who would normally be buying the Australian dollar, such as exporters hedging against a falling dollar, started to sell in order to profit from further falls in the Australian dollar (see Reserve Bank Bulletin (1999b)). Significant Reserve Bank intervention was required in early June 1998 to stop this aggressive selling from continuing (see Reserve Bank Bulletin (1998b)). While the respondents do not unanimously support speculation as a stabilising force, 91.2% of the respondents feel that speculation increased exchange rate volatility (see panel 2.12). This increased volatility can be partly explained by speculators building up and reversing profitable trading positions based on information that other market participants, such as fundamentalists, do not judge to be correct. Consequently, speculators adopt trading strategies differing from those of other market participants. For example, they could be buying or selling currencies at a time when other market participants are trading in the opposite direction or not trading very actively. The respondents also feel that speculative trading improves market efficiency and liquidity (see panel 2.12). Trading in a foreign exchange market is efficient when currencies trade at a value that fully reflects all the information available to market participants (see Blundall-Wignall et al. (1993)). Speculators can be seen as improving market efficiency because they will start trading when they perceive that current and expected market fundamentals have not been correctly priced into the existing exchange rates. They will then trade in a manner that will force the currency's value to change until it reaches its true value. This increase in efficiency will also be 19

accompanied by an increase in market liquidity as previously inactive dealers and dealers who normally trade for other reasons will be entering the market to trade for the purpose of speculation. 4.6 Intervention by the Reserve Bank Under a floating exchange rate regime central banks will intervene if they feel that exchange rate volatility has been excessive or exchange rates are under or overvalued. In Australia the timing and size of the intervention undertaken by the Reserve Bank has varied across several periods since the Australian dollar was floated in December 1983 (see Andrew and Broadbent (1994)). This variation has been due to a wide range of factors such as the severity of exchange rate volatility being experienced and how the intervention will affect other government policies. However, whatever the reason for the intervention most of the respondents feel that the Reserve Bank normally intervenes at the appropriate moment with only 14.3% feeling that its timing has been inappropriate (see panel 2.13). In recent years the Reserve Bank has seen the need to intervene several times. However, prior to 1997 the Reserve Bank had not intervened since November 1993. 72.7% of the respondents feel that Reserve Bank intervention has been successful in moving exchange rates towards their fundamental values. In fact, 76.55% of the respondents feel that Reserve Bank intervention achieves its goals while 23.5% do not. These responses support a study undertaken by Andrew and Broadbent (1994) on the effectiveness of Reserve Bank intervention. They argue that intervention will be successful if it is profitable because the Reserve Bank supports a depreciating Australian dollar by buying at low exchange rate levels and an appreciating Australian dollar by selling at high exchange rates. Their study found that the Reserve Bank's foreign exchange operations were profitable over most of the periods when intervention was taking place. While the Reserve Bank's trading on behalf of its clients, principally the Commonwealth Government, may have generated some of this profit, the continued finding of profitability over long periods does lend some support for the successfulness of intervention. However, the respondents are equally divided on the impact of Reserve Bank intervention on volatility with 50% feeling it decreases volatility and 50% feeling it 20

decreases volatility. This appears to be contradictory to the notion that successful intervention should be stabilising. However, the decreased volatility is more likely to occur over a longer term, as successful intervention begins to calm down market behavior (see Andrew and Broadbent (1994)). On the other hand intra-day volatility would increase when dealers attempt to trade out of their current market position if they had previously felt that the Reserve Bank was not going to intervene. 4.7 Time Horizons The ability to successfully predict exchange rate movements has been questioned by both academics and dealers. In panel 2.14 the respondents give very little support for exchange rate movements being either always predictable or never being predictable. The majority of respondents felt that if anything they were more likely to be sometimes predictable. They give slightly more support for intra-day exchange rate movements being sometimes predictable than they do for periods of less than six months and periods longer than six months. The type of factors that have a major influence on exchange rates differs depending on the time horizon in question. In panel 2.15 respondents claim that intra-day exchange rate movements are mainly determined by order placements followed by over-reaction to news, speculative forces, bandwagon effects and technical trading. The confidential nature of inter-dealer trades makes it very difficult to obtain data on the volume and price of individual trades. Consequently, the disclosure by respondents that order placements have a major influence on intra-day exchange rate movements indicates that studies of order flows should be able to reveal information on trading behavior. However, as the time horizon increases the respondents indicate that economic fundamentals have a growing impact on exchange rate movements while the other factors, apart from technical trading, become less significant, particularly over periods greater than six months. Technical analysis has the greatest influence on exchange rate movements up to six months but its impact is not as significant as the impact of fundamental analysis. 21

5. Conclusion This paper summarises the responses to a survey completed by dealers trading in the Australian foreign exchange market. Trading takes place at all hours of the day in this market with the market turnover and trading volume far exceeding that in other financial markets. Unfortunately, it has been difficult to develop models that can be used to accurately forecast exchange rate movements. As dealers are actively present in the market they will have experienced periods when exchange rates movements are considered to be normal as well as when they are considered to be irregular. So they are asked in the survey about their views on the factors influencing trading in the market. While the dealers do differ in their responses to several of the survey questions, a majority response was recorded for most of the questions. Consequently, the information gathered can provide insights to factors generating movements in exchange rates. The dealers surveyed indicate that their trading environment has changed over the last five years. A significant amount of trading is now being done through electronic broking rather than voice broking or directly with other dealers. Electronic broking makes it difficult for dealers to profit from quoting bid and asks which are different from those available under electronic broking. In fact the majority of dealers claim that they are now devoting more of their trading towards meeting client orders. Obviously they are finding it more profitable to provide services to their clients. This is supported by the large decrease over the last five years in dealers using jobbing as part of their approach to trading. Even with the change in their trading environment dealers do not feel that they are any closer to be able to accurately predict movements in exchange rates. In fact they see exchange rate movements as being only sometimes predictable. This unpredictability is not due to a lack of market competitiveness as dealers indicate that major players do not dominate trading in the market. In fact in order to maintain good trade relationships with each other dealers normally quote the same bid-ask spread size. This feeling of unpredictability has not prevented dealers from adopting trading strategies aimed at making a profit from exchange rate movements. The majority of 22

dealers adopt an even mixture of fundamental analysis and technical analysis. This mixture is required as changing economic fundamentals are only considered to affect exchange rates over a medium to long-term horizon. Although when announcements of Australian economic fundamentals are different from what the market expected, dealers will react within less than a minute to correctly price the anomaly into current exchange rates. The economic announcement having the biggest impact both today and five years ago is still interest rates. Excessive speculation and manipulation by hedge funds are seen to be the main factors that have prevented exchange rates from reflecting their fundamental value. Excessive intervention by central banks also received some support but dealers considered that intervention by Reserve Bank of Australia was generally successful and conducted at the most appropriate time. However, speculation cannot fully explain intra-day exchange rate movements. Dealers feel that order placements by clients and over-reaction of market participants to events are major influences on moment to moment movements in exchange rates. References Almeida A, Goodhart C and R. Payne (1998), "The Effects of Macroeconomic News' on High Frequency Exchange Rate Behaviour," Journal of Financial and Quantitative Analysis, 33(3), 383-408. Alreck P.L and R.B Settle, (1985) The Survey Research Handbook. Richard D.Irwin Andrew R and J Broadbent (1994) "Reserve Bank Operations in the Foreign Exchange Market: Effectiveness and Profitability." Research Discussion Paper 9406, The Reserve Bank of Australia. Arnott R.D and T.K Pham, (1993) "Tactical Currency Allocation." Financial Analysts Journal, September-October,pp 47-52. Anthony S., (1997) "Foreign Exchange in Practice." 2 nd edition, LBC Information Services. Banerjee A.V., (1992) "A Simple Model of Herd Behaviour." The Quarterly Journal of Economics, August, 3, pp 797-817. Bank of International Settlements, (1999) Central Bank Survey of Foreign Exchange and Derivatives Market Activity 1998, Basle, May. 23

Blundall-Wignall A, Fahrer J and A Heath, (1993)"Major Influences on the Australian Dollar Exchange Rate." From the proceedings of the Reserve Bank of Australia conference on The Exchange rate, International Trade and the Balance of Payments. Bollerslev T and M Melvin, (1994) "Bid-ask Spreads and Volatility in the Foreign Exchange Market." Journal of International Economics, 36, pp 355-372. Bullock M, Grenville S and G Heenan, (1993) "The Exchange Rate and the Current Account." From the proceedings of the Reserve Bank of Australia conference on The Exchange rate, International Trade and the Balance of Payments. Carew E., (1998) Fast Money 4. Australia: Allen and Unwin. Carew E and W Slayter,(1994) Forex: The Techniques of Foreign Exchange. 3 rd edition. Singapore: Heinemann Asia. Cheung Y and M.D Chinn, (1999a) "Traders, Market Microstructure and Exchange Rate Dynamics." Working Paper at Department of Economics, University of California, Santa Cruz. Cheung Y and M.D Chinn, (1999b) "Macroeconomic Implications of the Beliefs and Behaviour of Foreign Exchange Traders." Working Paper at Department of Economics, University of California, Santa Cruz. Cheung Y and C.Y Wong, (1999a) "Foreign Exchange Traders in Hong Kong, Tokyo, and Singapore: A Survey Study." Working Paper number 4256 at Department of Economics, University of California, Santa Cruz. Cheung Y and C.Y Wong, (1999b) "A Survey of Market Practitioners' Views on Exchange Rate Dynamics." Working Paper number 426 at Department of Economics, University of California, Santa Cruz. Chicago Mercantile Exchange, (1995) Futures and Options for Hedge Funds: The Regulatory Environment. Paper complied by the Chicago Mercantile Exchange. Gruen D and R.Kortian, (1996) "Why does the Australian Dollar Move so Closely with the Terms of Trade?" Research Discussion paper 9601, Reserve Bank of Australia. Huang R.D and R.W Masulis, (1999) "FX Spreads and Dealer Competition across the 24-Hour Trading Day." The Review of Financial Studies, Volume 12, number 1, pp 61-93. Kim S., (1998) "Do Australian and the US macroeconomic News Announcements affect the USD/AUD Exchange rate? Some Evidence from E-Garch Estimations." Journal of Multinational Financial Management, volume 8, issues 2-3, September. Krugman P.R.,(1989)"Exchange-Rate Instability." The Lionel Robbins Lectures, The MIT Press Cambridge. 24

Neely C.J., (1997) "Technical Analysis in the Foreign Exchange Market: a Layman s Guide." Federal Reserve Bank of St Louis Review, September/October, pp 23-38. Rankin B.,(1999) "The Impact of Hedge Funds on Financial Markets: Lessons from the Experience of Australia." Paper prepared for RBA Conference on Capital Flows and the International Financial System, Sydney, 9-10 August 1999. Reserve Bank of Australia, (1996) "Statement of the Conduct of Monetary Policy." Reserve Bank of Australia Bulletin, September, pp 1-3. Reserve Bank of Australia, (1998a) Survey of Foreign Exchange and OTC Derivatives Turnnover, September. Reserve Bank of Australia, (1998b) Reserve Bank of Australia Bulletin, August, pp 8. Reserve Bank of Australia, (1999a) Reserve Bank of Australia Bulletin, March, pp 9-12. Reserve Bank of Australia, (1999b) "Capital Flows and the International Financial System." Reserve Bank of Australia Bulletin, November, pp 46-54. Reserve Bank of Australia, (2000) "Inflation Trends and Prospects." Reserve Bank of Australia Bulletin, February, pp 29-35. Shleifer A and L.H.Summers, (1990) "The Noise Trader Approach to Finance." Journal of Economic Perspectives, vol. 4, no.2, Spring, pp 19-33. Singleton K.,(1987) "Speculation and the Volatility of Foreign Currency Exchange Rates." Carnegie-Rochester Conference Series on Public Policy, 26, pages 9-56. Amsterdam :North-Holland Taylor M., (1995)"The Economics of Exchange Rates." Journal of Economic Literature. Volume 33, March, pp 13-47. 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 304-314. 25

Appendix A: Survey Questions and Responses Section 1: Background Information Panel 1.1 Your current position is: % Treasurer/manager 38.5 Chief/senior dealer 53.8 Dealer/junior dealer 5.1 Other 2.6 Panel 1.2 Where are the headquarters for your foreign exchange products located? % Australia 33.3 Japan 20.5 Europe (excluding United Kingdom) 15.4 United Kingdom 10.3 Asia (excluding Japan) 10.3 United States of America 7.7 Canada 2.6 Panel 1.3 Your department's daily FX turnover (AUD$ million) on average is: % Below 100 43.6 100 to 500 28.2 500 to 1000 17.9 1000 to 5000 7.7 Over 5000 2.6 Panel 1.4 The percentage of your FX transactions that are traded via 8 : Interbank Traditional broker Electronic broker % % % 5 years ago Less than or equal to 20% 6.1 7.3 8.5 21% to 40% 14.6 11.0 1.2 41% to 60% 13.4 14.6 0.0 61% to 80% 8.5 9.8 1.2 81% to 100% 3.7 0.0 0.0 46.3 42.7 10.9 Today Less than or equal to 20% 13.0 15.2 9.8 21% to 40% 7.6 8.7 4.3 41% to 60% 4.3 4.3 4.3 61% to 80% 5.4 2.2 5.4 81% to 100% 12.0 1.1 2.2 42.3 31.5 26 8 The total of all responses for today and 5 years ago do not sum to 100% due to rounding. 26

Panel 1.5 The percentage of your FX transactions that are 9 : Interbank Business 10 Customer Business % % 5 years ago Less than or equal to 20% 3.9 10.5 21% to 40% 5.3 19.7 41% to 60% 18.4 15.8 61% to 80% 19.7 2.6 81% to 100% 2.6 1.3 Today Less than or equal to 20% 9.0 5.1 21% to 40% 10.3 20.5 41% to 60% 14.1 11.5 61% to 80% 14.1 11.5 81% to 100% 2.6 1.3 Panel 1.6 Your quotes with other Australian dealers are mainly for amounts (in AUD$): % Below 5 million 41.0 5 million 30.8 10 million 23.1 20 million 0.0 Over 20 million 5.1 Panel 1.7 Your quotes with other dealers operating in a different time zone are mainly for amounts (in AUD$): % Below 5 million 46.2 5 million 28.2 10 million 15.4 20 million 2.6 Over 20 million 7.7 Panel 1.8 The best way to describe your spot FX trading now and five years ago is that it: 5 years Now ago 11 % % Was driven by customer orders 35.5 45.6 Is an even mixture of technical analysis/fundamental analysis 27.4 26.3 Was the jobbing approach 22.6 10.5 Is mainly based on technical trading rules 6.5 10.5 Is mainly based on fundamental analysis 6.5 5.3 Other (please specify). 1.6 1.8 9 The total of all responses for today and 5 years ago do not sum to 100% due to rounding. 10 This includes trading to square customer transactions. 11 The total of all responses 5 years ago do not sum to 100% due to rounding. 27

Section 2: On the FX market Panel 2.1 The conventional bid-ask spread for each of the following exchange rates is 12 : Less than 3 points 3 to 5 points 5 to 7 points 7 to 9 points 9 to 10 points % % % % % EUR/USD 40.0 51.4 2.9 0.0 5.7 USD/JPY 30.6 63.9 0.0 0.0 5.6 AUD/USD 27.0 73.0 0.0 0.0 0.0 USD/DEM 26.3 68.4 0.0 0.0 5.3 NZD/USD 14.3 57.1 28.6 0.0 0.0 USD/GBP 5.9 64.7 20.6 2.9 5.9 USD/CHF 3.2 67.7 16.1 3.2 9.7 Panel 2.2 If most of your inter-bank quotation spreads conform to the market convention, please select the most important reason for such conformity 13 : % To maintain an equitable and reciprocal trading relationship with other traders 49.0 Your firm s policy 15.7 To secure a good market image for the firm 15.7 To maximise trading profits 11.8 To follow the practice of major players 7.8 Other (please specify) 0.0 Panel 2.3 Please indicate, for all inter-bank quotations, the proportion of your quotes that have a bid-ask spread larger or smaller than the market convention 14 : < 1% < 5% < 10% < 20% 20% % % % % % Larger than the convention 36.4 27.3 18.2 6.1 12.1 Smaller than the convention 38.7 9.7 16.1 6.5 29.0 Panel 2.4 Under most circumstances, the bid-ask spread of your inter-bank quote is mainly determined by 15 : % The type of customer you are dealing with 31.6 Market conditions (ie level of liquidity) 31.6 The market convention 24.6 The potential costs of making that quote 8.8 Other (please specify) 1.8 All of the above 1.8 12 The responses for each currency do not always total to 100% due to rounding. 13 A small number of respondents selected more than one reason. 14 A small number of respondents did not answer this question. 15 A small number of respondents selected more than one reason. 28

Panel 2.5 Please check the 3 (or fewer) most important reasons for you to quote a bidask spread different from the market convention: % A thin or hectic market 28.9 The customer is someone you do a lot of business with 20.6 Increased market volatility 14.4 Before and after the announcement of market news 12.4 An unexpected change in market activity 10.3 The customer is someone you do not want to do business with 8.2 Holding a position against the market trend 2.1 An increase in the costs of keeping the position 2.1 Other (please specify): 1.0 Panel 2.6 Do you agree that one or a few major financial institutions dominate trading in the following currencies in the Australasian region? Yes No No opinion % % % NZD/USD 56.4 35.9 7.7 USD/CHF 16 32.4 59.5 8.1 AUD/USD 30.8 66.7 2.6 USD/GBP 30.8 56.4 12.8 EUR/USD 20.0 80.0 11.4 USD/YEN 20.5 71.8 7.7 USD/DEM 17 16.1 67.7 16.1 Panel 2.7 Select the 3 (or fewer) most important sources of competitive advantage for the large players in the FX markets: % A large customer base 38.1 Better information about the market 22.7 Experienced traders 14.4 Ability to deal in large volumes 11.3 Ability to influence exchange rates 4.1 Ability to offer new FX products 4.1 Lower operating costs 3.1 Smaller counterpart risks 0.0 Other (please specify) 18 : 2.1 16 A small number of respondents did not respond. 17 A small number of respondents did not respond. While the euro replaced the deutsche mark at the start of 1999 for use in international trade it can still be used domestically until 1 January 2002 when national notes and coins will gradually be replaced with euro notes and coins. Deutsche marks have to be fully removed by 1 July 2002. 18 One respondent stated that major players had the ability to establish a strong franchise while another stated that they were subject to good credit limits. 29

Panel 2.8 In your opinion, how long does it take for dealers to trade after they find out that the following economic announcements from the major developed countries are different from what the market expected? 19 < 10 sec. < 1 min. < 10 min. < 30 min. > 30 min. % % % % % Inflation 74.4 25.6 0.0 0.0 0.0 Unemployment 69.2 30.8 0.0 0.0 0.0 Trade deficit 69.2 25.6 5.1 0.0 0.0 Current account 71.8 25.6 2.6 0.0 0.0 Interest rate 74.4 20.5 5.1 0.0 0.0 Retail sales 66.7 25.6 5.1 2.6 0.0 GDP (GNP) 69.2 25.6 5.1 0.0 0.0 Other (please specify) 20 : 50.0 50.0 0.0 0.0 0.0 Panel 2.9 Choose the economic announcement from the following counties that you consider has or had 5 years ago the biggest impact on FX markets in Australia? From Australia From USA 21 % % 5 years ago 22 Interest rate 27.5 28.8 Current account 17.5 9.1 Inflation 15.0 24.2 Trade deficit 13.8 10.6 Unemployment rate 12.5 10.6 GDP (GNP) 11.3 12.1 Retail sales 2.5 4.5 Other (please specify) 0.0 0.0 Today Interest rate 28.2 32.4 Inflation 17.6 25.4 Unemployment rate 16.5 18.3 Current account 11.8 2.8 Trade deficit 9.4 4.2 GDP (GNP) 9.4 9.9 Retail sales 7.1 7.0 Other (please specify): 0.0 0.0 19 The responses for each announcement do not always total to 100% due to rounding. 20 These were announcements made by the Federal Open Market Committee (FOMC) meeting in the United States of America and the Bank of Japan's Tankan announcement. 21 Respondents were given the option of choosing economic announcements from a foreign country other than the USA but none of the respondents took up this option. 22 The responses for 5 years ago do not total to 100% due to rounding. 30

Panel 2.10 Do you believe movements in an exchange rate accurately reflects changes in its fundamental value? Yes No No opinion % % % Intra-day 12.8 84.6 2.6 Medium run (within 6 months) 64.1 28.2 7.7 Long run (over 6 months) 84.6 10.3 5.1 Panel 2.11 If the FX market does not accurately reflect an exchange rate s fundamental value, which of the following factors are responsible for this? Yes No No opinion % % % Excessive speculation 84.2 7.0 7.9 Manipulation by hedge funds 81.1 8.1 10.8 Excessive central bank intervention 52.8 36.1 11.1 Views taken by major trading banks 36.4 51.5 12.1 Dealers being slow to respond 10.0 73.3 16.7 Other (please specify) 23 : 100.0 0.0 0.0 Panel 2.12 In your opinion, speculation mainly: Yes No % % Increases exchange rate volatility 91.2 8.8 Increases market liquidity 88.2 11.8 Improves market efficiency 71.9 28.1 Moves exchange rates towards their fundamental value 55.6 44.4 Panel 2.13 In your opinion, intervention by the Reserve bank of Australia: Yes No % % Are usually conducted at the appropriate moment 85.7 14.3 Achieve the desired goals 76.5 23.5 Moves exchange rates towards their fundamental value 72.7 27.3 Increases exchange rate volatility 50.0 50.0 Panels 2.14 Do you believe exchange rate movements are predictable? 24 Always Sometimes Never No opinion % % % % Intra-day 2.6 82.1 12.8 2.6 Medium run (within 6 months) 10.3 79.5 10.3 0.0 Long run (over 6 months) 15.4 69.2 15.4 0.0 23 Other reasons given were market sentiment, the longer time needed to complete large orders, view of fund managers, capital flows and technical breaks in a particularly long or short market. 24 The responses for each time horizon do not always total to 100% due to rounding. 31

Panel 2.15 Select the single 25 most important factor that determines exchange rate movements in each of the three time horizons 26 : Intra-day Medium run Long run % % % Order placements 28.1 1.8 0.0 Over-reaction to news 25.8 1.8 0.0 Speculative forces 20.2 19.3 2.9 Bandwagon effects 13.5 12.3 0.0 Technical trading 10.1 22.8 8.6 Economic fundamentals 1.1 40.4 88.6 Other (please specify) 27 : 1.1 1.8 0.0 25 A small number of respondents selected more than one factor for each time horizon. 26 The responses for each time horizon do not always total to 100% due to rounding. 27 Option positions were the other factor considered to have an impact on intra-day and medium run exchange rate movements. 32