Revised October 6, 19951
The entire document is available as a PDF file from the Bank for International Settlements.
Beginning informally in the 1960's, and evolving into a global market survey conducted simultaneously by 26 central banks, the Foreign Exchange Survey has been conducted every three years, most recently in April 1995. In coordination with other central banks, which are also releasing their national results at this time, this report describes the results of the United States survey conducted by the Federal Reserve Bank of New York. The Bank for International Settlements (BIS) will aggregate the data from each of the national surveys to produce global market statistics to be reported in a forthcoming BIS publication.
The focus of the Survey is the size and structure of the foreign exchange market, and its evolution over time. In contrast to previous surveys, the April 1995 survey also included financial derivatives on currencies, interest rates, equity markets, and some commodities. The results described below, however, cover only the results of the foreign exchange part of the Survey. The results of the financial derivatives section of the survey will be presented in a later publication.
The products included in this survey are: foreign exchange-spot; foreign exchange-forwards; and foreign exchange-swaps. (Cross-currency interest rate swaps/ currency swaps are not included here, but are in the derivatives survey.)
The foreign exchange market has continued to grow rapidly, though at a slower pace than during the 1980's.
In April 1995, turnover volume in the U.S. was 46% higher than in April 1992; an annualized growth rate of 13% (somewhat less than the 20% annualized growth rates seen in the 1980's). Average daily turnover volume of foreign exchange transactions in the U.S. was $244 billion in April 1995. (This figure has been adjusted for double counting of transactions between dealers in the survey see item (g) in the Definitions and Explanatory Notes.)
Currency pair highlights:
The U.S. dollar (USD), Deutsche Mark (DEM), and Japanese Yen (JPY) remain the most actively traded currencies, in that order, with total turnover in each roughly twice as large as the next largest.
The most traded currency pairs, by a large margin, were: USD/DEM with 30% of turnover volume; and USD/JPY with 20% of turnover volume.
A notable development since the last survey was the rapid growth of transactions in the French Franc, other European Union, and emerging market currencies. (Other European Union currencies are those other than the Deutsche Mark, the British Pound, and the French Franc.) Turnover volume in these currencies grew at a rate almost three times higher than the overall market growth.
The growth of transactions in these currencies has created demand for greater use of direct trading in non-U.S. dollar trades transactions that do not go through the U.S. dollar as a vehicle currency. This development is evident in the higher share (to 13% from 10%) of Deutsche Mark/non-U.S. dollar trades (DEM transactions excluding USD/DEM trades), and the fall in the share of U.S. dollar/Deutsche Mark trades (to 30% from 33%).
Competition.The market structure statistics suggest that the foreign exchange market is highly competitive. Among the top ten dealers, only four dealers' ranking remained unchanged between 1992 and 1995. Among the dealers who were in the top ten in either 1992 or 1995, four dealers saw their ranking fall by five or more places, while four dealers saw their ranking rise by five places or more.
Electronic trading. The use of automated brokerage (electronic order matching) has grown rapidly from virtually nothing in 1992. Almost a third of brokered spot transactions, and 6% of total market volume was conducted through automated brokerage systems.
They also trade in the afternoon. The percentage of volume traded in the afternoon (12pm - 4pm) has grown (to 34% from 29% in 1992). Most trading still occurs in the morning, however, with almost twice as much volume in the morning than the afternoon (58% between 8am and 12pm, and 34% between 12pm and 4pm).
Cross border transactions. Transactions between a dealer located in the U.S. and a counterparty abroad accounted for more than half of turnover volume (58% of turnover). Since most of the U.S. trading occurred in the morning, much of the cross-border volume was probably with counterparties in Europe.
Counterparty types. Most trading was between financial counterparties (83% of volume). However, the percentage of trading volume with non-financial customers has grown (to 17% from 14% in 1992).
DEFINITIONS AND EXPLANATORY NOTES
Additional discussion of the survey method and definition of terms can be found in Annex I.
- This report provides information about the over-the-counter foreign exchange market, where virtually all wholesale foreign exchange transactions take place.
- Foreign exchange related transactions on organized exchanges (futures exchanges) account for only a small fraction of total currency related trading (less than 6 percent of total volume in the 1992 Survey). Figures for exchange-traded contracts are not included in this report, but will be included in a forthcoming report on the financial derivatives survey.
- Other currency products (FX futures, FX options, and cross currency interest rate swaps/currency swaps) are in the financial derivatives survey and will be reported at a future date.
- The 1992 Survey included FX futures and FX options which amounted to 13% of the 1992 turnover volume. For comparability, the 1992 turnover figures used in this report were recalculated to exclude options and futures. For this reason, the 1992 figures in this report will be about 13% smaller than the turnover totals reported in the 1992 survey report.
- All transactions in the foreign exchange market are conducted with a foreign exchange dealer as a counterparty, so that the data provided by the dealers capture activity in the entire market. Unless otherwise stated, the market figures in this report were produced from the dealers' data; figures that were produced from the brokers' data are identified as brokers' market figures.
- Some transactions in the foreign exchange market are arranged by a broker who matches the bid and ask quotes of interbank traders. While trades arranged in the brokers' market are already included in the market totals obtained from the dealers' data, the brokers' market was surveyed separately in order to obtain a more complete picture of the brokers' market (a separate survey of the brokers market also provides an independent check of trends in the dealers' data).
- Adjusted figures are market totals after adjusting for double counting while unadjusted figures are gross totals without adjusting for double counting. The results reported are based on adjusted figures unless otherwise stated.
1. Turnover volume
- The foreign exchange market has continued to grow rapidly, though at a slower pace than during the 1980's (see Chart 1).
- In April 1995, turnover volume of foreign exchange transactions in the United States was 46% higher than three years ago; an annualized growth rate of 13% (somewhat less than the 20% annualized growth rate of the 1980's).
- In April 1995, average daily turnover volume of foreign exchange transactions in the United States was $244 billion. (This figure was obtained after adjusting for double counting of transactions between dealers in the survey; the unadjusted total was $295 billion see the Methodology section for further details.)
- Turnover through the brokers' market has fallen to 30% of total turnover in 1995 from 35% in 1992, and from about 50% in the 1980's.
- Average daily turnover of transactions in the brokers' market was $73 billion in April 1995, about 27% higher than in April 1992 with most of the growth (or 82%) occurring in automated brokerage systems (see Chart 8).
2. Which Currencies were most actively traded?
- The U.S. dollar (USD), German Deutsche Mark (DEM), Japanese Yen (JPY), British Pound (GBP), Swiss Franc (CHF), and French Franc (FRF) were the most actively traded currencies (see Chart 3).
- The most traded currency pairs, by a large margin, were: USD/DEM with 30% of turnover volume; and USD/JPY with 20% of turnover volume.
- Virtually all foreign exchange trades involved either the USD, DEM, or JPY. Transactions in which neither leg of the trade involved one of these three currencies amounted to only 1% of total turnover.
- The USD, DEM, and JPY remain the most actively traded currencies, in that order, with total turnover volume in each twice as large as the next largest. The multiple-of-two relationship in the relative volume in these currencies was also present in the 1992 Survey (both in the U.S. and globally), but not in the 1989 Survey.
- The sum of transactions in which one leg was the:
- USD amounted to $209 billion daily, or 86% of total turnover;
- DEM amounted to $104 billion daily, or 43% of total turnover;
- JPY amounted to $55 billion daily, or 23% of turnover; where the sum will exceed 100% because the figure for each currency includes transactions with the others. (For example, a USD/DEM transaction appears in both the USD total and the DEM total.)
3. Currency trends
- A notable development since the last survey was the growth of volume in the French Franc (FRF), other European Union, and emerging market currencies. Transaction volume in these currencies more than doubled a growth rate almost three times higher than the overall market growth (see Charts 4a and 4b). (Other European Union currencies are those currencies other than DEM, GBP, and FRF)
- Turnover volume in the French Franc, other European Union, and emerging market currencies:
- increased by 130% against the USD;
- increased by 143% against the DEM.
- The share of these currencies in total turnover is now 22%.
- Table 2 lists the currencies in the "other" category that were traded in significant volume.
- This change in trading patterns can be seen from the data in Chart 4a, with the following partition of the currency pairs, where the share of:
- DEM/non-USD transactions rose (to 13% from 10%);
- USD/DEM transactions fell (to 30% from 33%);
- the share of USD/non-DEM transactions remained unchanged at 57%.
- The smaller share of the USD was a trend present in the 1992 survey, and is also apparent in the 1995 Brokers' data.
- Among non-USD transactions, the DEM/French Franc was the most actively traded currency pair, with average daily turnover of $8.5 billion, or 3.5% of turnover volume.
- The share of the DEM in the brokers' market increased by a large amount to 60% from 47% in 1992 (see Charts 5a and 5b).
- The DEM is now a close second to the USD in the brokers' market. For example, the sum of transactions in the brokers' market in which one leg was the:
- USD amounted to 82% of brokers' turnover;
- DEM amounted to 60% of brokers' turnover;
- JPY amounted to 22% of brokers' turnover; where the sum exceeds 100% because the figures for each currency includes transactions with the others.
- The increase in the volume of brokered DEM transactions occurred against almost all currencies. The growth in transactions between the DEM and currencies other than the USD was most notable, with an increase to 18% of brokered transactions from 11% in 1992.
- The decline of the share of brokered USD transactions was due to very little growth in volume of brokered transactions between the USD and currencies other than the DEM. For example, the volume of USD/JPY transactions in the brokers' market remained constant at $13 billion daily.
- This development is apparent in the following partition of the currency pairs in the brokers' market (see Chart 5a), where the share of:
- DEM/non-USD trades rose (to 18% from 11%);
- USD/non-DEM transactions fell (to 40% from 52%);
- USD/DEM trades rose (to 42% from 37%).
4. Foreign exchange products
- As in 1992, turnover volume was concentrated in spot transactions (55%), followed by FX swaps (34%), and FX forwards (11%) see Chart 7. (Note: cross-currency interest rate swaps/currency swaps are not included in the category FX swaps.)
- While total volume of spot transactions increased, the share of spot transactions in total volume fell slightly (by 2 percentage points) between 1992 and 1995. This rise in the proportion of forwards and swaps is a continuation of a trend present in earlier surveys.
- The disaggregated figures for forwards and swaps may be unreliable, because (in both the 1992 and 1995 surveys) a few firms whose internal databases combined forward and swap transactions had difficulty disaggregating the two products for the survey. In any event, the sum of forwards and swaps can be taken with confidence.
- up to seven days, 66%;
- over seven days to one year, 32%
- over one year, 2%.
- This increase in the share of spot transactions in the brokers' market was due to:
- the rise of automated brokerage systems which still broker only spot transactions;
- a decline in brokered FX swap transactions.
5. Market Structure
- The market structure statistics suggest that the foreign exchange market is highly competitive. The most notable development in support of this inference was the change in the composition of the set of largest dealers.
- Among the top ten dealers, only four dealers' rankings remained unchanged between 1992 and 1995, and three of the top ten were not among the top ten in 1992.
- Among the top ten dealers in either 1992 or 1995, four dealers saw their ranking fall by five places or more, while four dealers saw their ranking rise by five or more places.
- While the market shares of the larger dealers rose, the highest growth rate of market shares was in the group between the top ten and top 20 dealers (to 23% from 19%).
- The cumulative market shares of the largest dealers in the survey were:
- the cumulative share of the largest five: 30% (26% in 1992);
- the cumulative share of the largest ten: 47% (41% in 1992);
- the cumulative share of the largest 20: 70% (60% in 1992);
- the top half of dealers had a cumulative share of: 96% (95% in 1992).
- Most trading was between financial counterparties (83% of volume). However, the percentage of trading volume with non-financial customers has grown (to 17% from 14% in 1992).
- The share of interdealer transactions in total turnover fell (to 56% from 68% in 1992) see Chart 9.
- While the volume of interdealer transactions rose, the volume of transactions between dealers and customers grew at an even faster pace. Transaction volume of:
- interdealer trades grew by 22%;
- trades with non-financial customers grew by 78%;
- trades with financial customers rose by 100%.
Some of this change could be attributed to the twenty small firms who participated in the 1992 survey but did not in the 1995 survey. If these firms are still actively trading, then dealers' trades with these firms would appear in the "financial customers" category. These firms accounted for 7.5% of volume in the 1992 survey but probably a smaller share of turnover in 1995, if we infer from their dropping out of the survey that they were less active in the market. In any event, even if the larger estimate of the missing firms' share of volume (7.5%) were used to adjust the 1995 counterparty breakdowns, the fall in the inter-dealer share of volume would still be present.
- Notably, the high proportion of cross-border activity was conducted across all types of counterparties: inter-dealer, other financial customers, and non-financial customers (see Chart 10).
- Relative to 1992, the proportion of cross-border activity in the interdealer market increased slightly, to 63% of interdealer volume from 60% in 1992.
- Among interdealer trades, spot amounted to 60% of interdealer volume.
- Among trades with non-financial customers, forwards and swaps amounted to 53% of volume with non-financial customers (50% in the case of financial customers).
Looking at the data from a different perspective, the majority of forwards volume was traded with customers (financial and non-financial), while the majority of spot volume was traded with other dealers, and swaps were about evenly split between dealers and customers.
- In the forward market, customer trades amounted to 63% of forward volume.
- In the spot market, interdealer trades amounted to 61% of spot volume.
- Among the top twenty dealers, the dealers whose ranking rose between 1992 and 1995 had a slightly higher credit rating, on average, than the average credit rating of the dealers whose ranking fell.
- The share of non-banks in turnover volume of the top twenty dealers remained comparable to their 1992 share. The share of non-banks in the turnover volume of the top ten dealers fell (to 15% from 20%), but their share of the top twenty firms' volume rose (to 24% from 21%).
5. Trading patterns
Electronic trading systems
- The use of automated brokerage (electronic order matching) has grown rapidly, to almost a third of brokered spot transactions from almost nothing in 1992 (see Charts 11 and 12).
- The volume of turnover conducted through automated brokerage amounted to:
- 6% of total market volume;
- 19% of volume in the brokers' market;
- 29% of brokered spot volume.
Activity by time of day
- While more than half of turnover volume was transacted in the morning, the percentage of volume traded in the afternoon has grown.
- Almost twice as much volume was conducted in the morning than the afternoon (58% between 8am and 12pm, and 34% between 12pm and 4pm).
- The percentage of volume traded in the afternoon (12pm - 4pm) has grown (to 34% from 29% in 1992) see Chart 13.
- In spite of the global nature of the foreign exchange market, trading activity follows the path of sunlight around the globe. In the United States, only 8% of daily turnover occurs overnight (4pm - 8am) but that share has grown from 5% in 1992.
- The bottom half of the dealer population (65 firms with a market share of about 4%) did 65% of their turnover in the morning as opposed to 58% for the market as a whole.
Average Deal Size
- The average size of a deal in the market as a whole remained comparable to 1992.
- The average size of a transaction was largest for FX-swaps at $15 million per trade, while spot transactions were much smaller at $3.5 million per trade.
- In the brokers' market, the average size of forward and swap trades were much larger than in the market as a whole as was the case in 1992.
1. This revision corrects figures relating to the brokers' market. The results relating to the overall market remain unchanged, and the overall market figures are identical to those in the initial report.