EKATERINA E. EMM GERALD D. GAY*

Transcription

1 THE GLOBAL MARKET FOR OTC DERIVATIVES: AN ANALYSIS OF DEALER HOLDINGS EKATERINA E. EMM GERALD D. GAY* We provide a descriptive examination of the trading activities of one of the most important intermediaries in global financial markets the OTC derivatives dealer. These dealers play a central role in the provision of derivative products and in the intermediation of market risks faced by financial and nonfinancial firms alike. Utilizing a unique database, we analyze the derivatives holdings of 264 dealers spanning 34 countries over the period We document the geographic composition of dealers on both country and regional levels as well as analyze trends in dealer holdings on an aggregate and individual product level. We further analyze the extent of global merger activity among dealers and resulting consolidation effects. Finally, we investigate at the individual dealer level The authors gratefully acknowledge the helpful comments and suggestions of Milind Shrikhande, Ufuk Ince, and Anna Agapova along with those of an anonymous reviewer. We have benefited from discussions with Andrei Osonenko of Swaps Monitor Publications regarding the data used in this analysis. *Correspondence author, Department of Finance, J. Mack Robinson College of Business, Georgia State University, MSC 4A1264, Gilmer Street SE, Unit 4, Atlanta, GA 30303; ggay@gsu.edu Received September 2003; Accepted February 2004 Ekaterina E. Emm is Genevieve Albers Visiting Fellow at Seattle University in Seattle, Washington. Gerald D. Gay is a Professor and Chairman in the Department of Finance at J. Mack Robinson College of Business at Georgia State University in Atlanta, Georgia. The Journal of Futures Markets, Vol. 25, No. 1, (2005) 2005 Wiley Periodicals, Inc. Published online in Wiley InterScience ( DOI: /fut.20138

2 40 Emm and Gay the extent and evolution of their array of product offerings Wiley Periodicals, Inc. Jrl Fut Mark 25:39 77, 2005 INTRODUCTION In this article we provide a descriptive look at the trading activities of one of the most important intermediaries in global financial markets the over-the-counter (OTC) derivatives dealer. The proliferation and use of derivatives during the past two decades, especially those traded OTC, has been among the most spectacular developments in financial markets. According to recent estimates by the Bank of International Settlements (2003), as of year-end 2002 the global outstanding notional amount of OTC derivatives (e.g., swaps, forwards, and options) had grown to over $141.7 trillion. 1 Clearly, the growth in this market has been driven by the needs of firms seeking risk management solutions and the ability of financial engineers and dealers to respond. This study furthers our understanding of the structure of this important market and complements an existing literature that has largely focused on the demand for hedging services. In particular, an extensive literature has emerged that analyzes rationales for hedging and the demand for derivatives, typically presented from the perspective of the corporate end user. 2 In contrast, the supply side has received little attention, especially in regards to the activities of OTC derivatives dealers who play a central role in the provision of derivative products and in the intermediation of market risks faced by financial and nonfinancial firms alike. 3 The study is also important in that it addresses the extent to which this market has evolved along international dimensions and in specific product offerings. 1 By comparison, the BIS also reports the global exchange-traded derivatives market at $23.9 trillion notional outstanding as of year-end A comparison of the market structures of OTC and futures exchange trading is discussed in Kamara (1988) and Stulz (2003). 2 For reviews of this literature see, for example, Smithson (1998), Allayannis and Ofek (2001) and the seminal papers of Smith and Stulz (1985), Froot, Scharfstein, and Stein (1993) and DeMarzo and Duffie (1995). Also, Bartram, Brown, and Fehle (2003) provide a recent examination of the hedging practices and determinants of derivatives usage of a comprehensive and global sample of nonnonfinancial firms. 3 Studies by Sun, Sundaresan, and Wang (1993), Kambhu, Keane, and Benadon (1996), and Malhotra (1997) discuss the role of derivatives dealers in facilitating the passage of price risk from end users to other market participants. Dodd (2002) describes various organizational forms along which derivatives dealers provide intermediation services. These include traditional dealer markets wherein bids and offers are quoted (often orally over the telephone or on electronic bulletin boards) with transactions negotiated on a bilateral basis, as well as electronic trading platforms in which bids and offers are posted and trades are executed against these quotes.

3 Global Market for OTC Derivatives 41 To these ends we analyze the derivatives holdings of 264 dealers spanning 34 countries over the period Our analysis uses a unique database obtained from Swaps Monitor Publications, Inc. This Database of Dealer Outstandings provides what we understand to be the most comprehensive collection of disaggregated holdings information of OTC derivatives dealers and thus permits identification of specific dealers and their derivatives holdings on a longitudinal basis. 4 Position information includes not only total derivative holdings (measured in notional dollars outstanding), but in most instances asset group breakdowns for interest rate, currency, equity, commodity, and credit derivatives. For many dealers, further breakdowns within these asset groups are reported for specific product lines including swaps, forwards, and option contracts. Our investigation begins with a global perspective on the size of this market and trends that have occurred in both aggregate holdings and subgroupings over the period. We find that the overall OTC derivatives market has grown substantially, having more than doubled in size from our initial 1995 reported numbers. The growth is primarily driven by trade in interest rate derivatives, which make up roughly threefourths of the market. However, despite observing growth in most of the other derivative asset groups, trade in currency derivatives has been relatively flat with total positions actually somewhat below peak levels of 1997, a pattern we attribute to the consolidation of several currencies into the euro. With respect to specific product lines, we find that swaps are overall the most prevalent interest rate derivative. However, forward contracts are the leading instrument among the currency derivatives, whereas options lead in the case of both equity and commodity derivatives. Next, we document the distribution of dealers by country and by geographic region (e.g., North America, Europe, and Asia/Pacific) and make comparisons on the basis of the number of active dealers and global market share. We find that although the United States is the leading country in terms of both the number of dealers and global market share, there is substantial activity conducted by dealers based in Germany, Japan, Britain, and France. In fact, on a geographic basis, the 4 The only other data source known to us that is similar in this regard is The Office of Comptroller of Currency s (OCC) Bank Derivatives Report, which is based on information collected from Reports of Condition and Income (call reports). The main drawback of OCC data set, however, is that it is limited to banking organizations in the U.S.

4 42 Emm and Gay European region has the largest contingent of dealers and market share. Using individual dealer market shares, we compute levels of and show trends in dealer concentration on both a global level and within the United States. On both levels, dealer concentration has grown significantly over our sample period. For example, the 4- and 20-firm concentration ratios for the global population of dealers have risen from 14% to 28% and 48% to 67%, respectively. In the United States, the statistics are even higher, with ratios having risen to 69% and 98%, respectively. We briefly discuss potential concerns that these statistics may suggest regarding systemic risk in the financial system. We also analyze the extent of global merger activity that has occurred among derivatives dealers and discuss various effects on industry structure. Our final set of analyses focuses on the extent of asset group and product line offerings of dealers. Attesting to their market breadth, the number of dealers making markets in the various asset groups is large and is greatest in interest rate and currency derivatives (an average of 197 and 201 dealers, respectively). This is followed by equity derivatives (116 dealers), commodity derivatives (66 dealers), and credit derivatives (27 dealers). Across each of the North American, European, and Asian/Pacific regions, comparable percentages of dealers make markets in interest rate and currency derivatives (more than 90% of dealers in each region) and equity derivatives (more than 50% participation). However, significant regional differences exist in commodity and credit derivatives, with North American and Asian/Pacific dealers being the most active, and European dealers the least active. Consistent with these findings, dealers in the North America and the Asia/Pacific regions offer the largest array of individual product lines. The rest of this article is organized as follows: the next section, Database Description, describes the data set on which we base our analysis. Aggregate Dealer Holdings provides statistics on the size and trends of aggregate dealer holdings along with breakdowns for holdings in various derivative asset groups and product lines. Geographic Composition and Concentration Levels documents the geographic composition of dealer activity along with concentration levels, while Merger Activity presents our merger analysis. In Extent of Asset Group and Product Line Participation, we analyze, at the individual dealer level, the extent of participation in each of the various asset groupings and product lines. The last section provides concluding remarks.

5 Global Market for OTC Derivatives 43 DATABASE DESCRIPTION Our study uses the June 2003 edition of the Database of Dealer Outstandings, published by Swaps Monitor, Inc. 5 Our analysis focuses on all dealers reported in the database disclosing derivatives holdings information over the period. This entails 264 dealers spanning 34 countries. Swaps Monitor states that their database consists of all reliably accurate data that have been publicly disclosed by all dealers since Sources used by Swaps Monitor to derive derivatives holdings include audited financial statements, regulatory filings, reports to shareholders, or other documents subject to similar standards of accuracy. To be considered a dealer for inclusion in the database, at least one of three criteria must be met: The firm is a primary member of the International Swaps and Derivatives Association (ISDA) the firm has total derivatives for trading purposes of at least $10 billion; or the firm has commodity or equity derivatives for trading purposes of at least $1 billion. All positions are reported in notional U.S. dollar amounts with non-u.s. dollar holdings converted to dollars at the prevailing exchange rate as of the balance sheet date. 6 For reference to the interested reader, we present in Appendix A an alphabetic listing of all dealers that appear in the database for the period. For each dealer, we identify their country of origin and all years for which position information is reported. In addition, we indicate those years when a dealer ceased reporting because of being acquired by or merging with another dealer (denoted by M), and in the last column we identify the acquiring firm. We also note those dealers and years in which Swaps Monitor has indicated that either disclosure 5 Swaps Monitor Publications, Inc. is a private company located at 401 Broadway, Suite 610, New York, NY 10013; telephone Until 1997 the firm published its annual Database of Users of Derivatives, which focused on the derivatives holdings of end-users. The firm has since continued to serve as a leading industry vendor of derivatives data focusing primarily on the provision of quantitative information regarding the activities of derivatives dealers. 6 A number of entities are now conducting regular surveys of various segments of the OTC derivatives markets. Unfortunately, information on individual dealer holdings is not made available as surveyed information is kept on a confidential basis with only aggregated information reported. Among the more comprehensive surveys are those conducted by the International Swaps and Derivatives Association (ISDA), the Bank for International Settlements (BIS), and the British Bankers Association (BBA). Initiated in 1989, the ISDA Market Survey is conducted semiannually and covers the holdings of their primary membership. The BIS publishes 2 surveys: The Regular OTC Derivatives Market Statistics, which has been conducted semiannually since June 1998, and The Triennial Central Bank Survey of Foreign Exchange and Derivatives Market Activity which has been published triennially since Both surveys are based on information collected by the central banks of the G10 countries on major banks and dealers. Finally, the BBA has conducted their Credit Derivatives Survey every two years since 1996.

6 44 Emm and Gay has ceased or is not available (denoted by DC). These omissions occur mainly in 2001 and were caused, in large part, by changes in reporting requirements. The implementation of Financial Accounting Standard (FAS) 133 and revisions to International Accounting Standard (IAS) 39 led many firms to report market or replacement values rather than notional holdings. 7 Finally, we indicate any change in name that a dealer may have experienced. Swaps Monitor reports dealers OTC holdings on three levels. First, for each dealer, information is provided on their total notional outstandings for all derivatives positions combined. Second, for most dealers, Swaps Monitor provides breakdowns of total dealer holdings in each of five asset groups: interest rate, currency, equity, commodity, and credit derivatives. Third, for about three-fourths of all dealers, further breakdowns of these asset group holdings are provided along three product lines: swaps, forwards, and options (with the exception of credit derivatives). While acknowledging that our data set does not contain an exhaustive listing and breakdown of all derivatives holdings of all dealers, we still report and discuss our findings as though it captures a reasonable representation of the global OTC dealer market. AGGREGATE DEALER HOLDINGS To provide context to our subsequent analysis, we begin by presenting the reader with an annual summary of global dealer holdings with breakdowns for each derivative asset group and, within each group, each product line. 8 These annual totals for the period , reported in terms of notional dollar holdings, are provided in Table I. Before commenting on the findings, we note that total dealer holdings will exceed total actual outstanding positions resulting from interdealer transactions. That is, when a dealer s counterparty is another dealer, as opposed to an end user, the derivative will show up on the books of both dealers. Although our database does not provide information about counterparties, statistics reported in various editions of the ISDA Operations Benchmarking Survey indicate that dealers consider 35% to 40% of their customers to be professional counterparties (e.g., other dealers). Still, 7 To ascertain the potential magnitude of these 2001 omissions on our 2001 analysis, we computed the year 2000 aggregate market share of these same dealers. This computed to be 8.9% of the global total of which 6.1% was attributed to two dealers, Goldman Sachs and Morgan Stanley. 8 As discussed above, because of incomplete disclosure by some dealers, totals for product lines within an asset group will not necessarily equal the total reported for that asset group, nor will totals across asset groups equal the total reported outstanding.

7 TABLE I Global Dealer Holdings of OTC Derivatives: (Notional Amounts in Millions of U.S. Dollars) Panel (a): Aggregate holdings by asset group and corresponding product lines Total Derivatives 77,506,608 92,519, ,058, ,212, ,274, ,470, ,177,229 Interest rate derivatives 40,684,054 53,696,915 67,736,797 95,239, ,981, ,292, ,677,847 Interest rate swaps 23,759,432 32,774,316 42,023,401 62,806,422 71,974,329 85,513,984 89,540,172 Interest rate forwards 5,691,174 6,237,371 7,914,908 9,072,842 11,152,715 9,989,006 11,336,604 Interest rate options 5,943,583 8,411,520 9,881,377 14,138,893 15,016,579 14,501,067 15,954,142 Currency derivatives 26,127,605 30,964,206 35,140,466 34,232,223 26,288,986 27,252,407 28,213,563 Currency swaps 2,016,376 2,522,180 2,788,281 3,406,985 3,842,552 4,714,237 6,391,315 Currency forwards 18,074,372 20,480,435 20,356,345 20,684,292 15,840,683 15,468,354 15,227,129 Currency options 2,510,345 3,805,983 5,706,037 4,965,583 3,167,199 3,320,318 3,533,331 Equity derivatives 799, ,868 1,366,470 2,427,094 2,655,399 3,029,986 2,367,393 Equity swaps 39,682 53,979 88, , , , ,759 Equity forwards 15,423 17,733 35,992 51, ,047 51,242 56,894 Equity options 640, ,335 1,030,517 1,982,536 2,049,303 2,516,334 1,893,702 Commodity derivatives 457, , , ,062 1,082,630 2,933, ,963 Commodity swaps 27,606 41,617 44,132 82,569 95, , ,852 Commodity forwards 121, , , , , ,785 71,150 Commodity options 91,237 94, , , , , ,664 Credit derivatives 0 17,538 54, , , , ,191 Panel (b): Aggregate product line holdings All swaps 25,843,096 35,392,092 44,944,139 66,415,071 76,081,598 90,725,970 96,319,098 All forwards 23,902,834 26,875,878 28,448,741 29,959,145 27,323,718 25,620,387 26,691,777 All options 9,185,891 12,968,152 16,725,841 21,246,000 20,437,630 20,699,270 21,569,839 Note. The table presents aggregate position holdings of OTC derivatives dealers. Total holdings as well as breakdowns for various asset groups and product lines are provided. The sample includes all dealers reported in the Database of Dealer Outstandings published by Swaps Monitor Publications, Inc.

8 46 Emm and Gay for purposes of our analysis and comments on industry structure, it is important to account for all dealer holdings. As reported in Panel A of Table I, we see that over reported global dealer holdings grew from $77.5 trillion to over $163 trillion over the 1995 through 2001 period. The largest asset group was interest rate derivatives, which, for example, in 2001 exceeded $120 trillion and composed about 74% of the global total. Currency derivatives composed the second largest group with $28 trillion in dealer holdings as of year-end 2001, or 17%. However, the growth in currency derivatives (aside from currency swaps) has not kept pace with that of other groups. To illustrate, in 1997, dealer holdings of currency derivatives peaked in excess of $35 trillion and were about 32% of the global total at that time. By 1999 the total had fallen to $26.2 trillion (18%). One reason for this decline is the introduction of the euro, which has reduced trading in a number of derivatives based on the former individual currencies of the various European Union countries. 9 The third largest asset group was equity derivatives, followed by commodity and credit derivatives. In Panel B we present the cumulative yearly totals for each of the various product lines (e.g., swaps, forwards, and options) after summing their respective totals in each asset group. Swaps have the largest annual totals and have been the most rapidly growing product line. In 2001, dealers reported swaps totals of $96.3 trillion, or approximately 60%, of all dealer holdings. This was followed by forward contracts, with reported holdings of $26.7 trillion, and options, with $21.6 trillion. Although swaps, in general, were the most popular overall product line, there were notable exceptions within the various asset groups. As shown in Panel A, among currency derivatives, forward contracts were consistently the most favored product line, and in some years, even currency options had larger totals than currency swaps. Further, in the equity and commodity groups, option totals typically exceeded both those of their swap and forward contract counterparts. GEOGRAPHIC COMPOSITION AND CONCENTRATION LEVELS We next inspect the global layout of dealer operations by first looking at the number of dealers headquartered in each country as well as aggregate country-level holdings relative to global totals. We then report global 9 A similar decline was observed during this time frame in the volume of exchange-traded currency futures at the Chicago Mercantile Exchange and which also has been attributed to the advent of the euro. For a discussion, see Back to the Futures in Chicago, Business Week Online, July 14, 2003.

9 Global Market for OTC Derivatives 47 and U.S. concentration levels based on holdings of individual dealers. We also note that dealer activity levels are attributed to the home country of the dealer, although many dealers will conduct trades through offices in other countries. 10 Geographic Distribution and Market Share Table II provides a look at dealer activity by country (and geographic region) over the period The table is presented in three panels: In Panel A we list the number of dealers in each country by year; 11 in Panel B we report the percentage of global derivatives holdings of all dealers in each country by year; and in Panel C we present a rank order comparison of country dealer activity using each of the two measures. In reviewing Panel A, we find that the United States leads all countries in terms of having the largest number of dealers with annual totals typically ranging from the mid- to upper 40s to a peak of 49 dealers in 1997 and Following the United States with the most dealers are Germany and Japan, each having comparable totals (mid-20s). These countries are then followed by Britain and France, which are also both comparable in dealer totals (mid- to lower teens). As shown in the bottom row of Panel A, the total number of dealers in each year has generally been in excess of 200, with a peak number of dealers occurring in 1997 when 237 dealers reported. By region, Europe has the greatest number of dealers with roughly 50% of the global total and approximately twice the total number of dealers found in North America (i.e., United States and Canada). Although led by Germany, France, and Great Britain, the dominance of the European region in terms of numbers of dealers is also a result of several other countries such as Italy, Switzerland, Belgium, Denmark, and Austria having a significant number of dealers. Within the Asia/Pacific region, Japan has the greatest number of dealers, followed distantly by Australia and Singapore. Initially surprising to us was the absence of dealers based in Hong Kong. Though we understand that a significant amount of derivatives activity takes place in Hong Kong, it appears to be originated through the local-based 10 To illustrate, our referee cites the case of Canadian Imperial Bank of Commerce. While the dealer is based in Toronto, Canada, it conducts many of its trades in London. 11 As discussed earlier, a number of U.S. and European dealers did not report notional holdings in 2001, thus causing 2001 to be an exception in parts of our analysis. We have repeated much of our analysis both with and without the 2001 data and obtain, except where otherwise noted, qualitatively similar findings.

10 48 Emm and Gay TABLE II Geographic Composition of OTC Derivatives Dealers: Country Panel A: Number of dealers by country and geographic region North America Canada US Total Europe Austria Belgium Britain Czech Denmark Finland France Germany Greece Ireland Israel Italy Netherlands Norway Poland Portugal Russia Spain Sweden Switzerland Total Asia/Pacific Australia China Hong Kong Japan Korea Malaysia Singapore Thailand Total Other countries Bahrain Brazil Saudi Arabia South Africa Total Grand total

11 Global Market for OTC Derivatives 49 TABLE II (Continued) Country Panel B: Global market share by country and geographic region (in percentage) North America Canada 3.79 % 3.56 % 3.34 % 2.99 % 2.81 % 2.56 % 3.09 % US Total Europe Austria Belgium Britain Czech Denmark Finland France Germany Greece Ireland Israel Italy Netherlands Norway Poland Portugal Russia 0.00 Spain Sweden Switzerland Total Asia/Pacific Australia China Hong Kong 0.03 Japan Korea Malaysia Singapore Thailand Total Other countries Bahrain Brazil Saudi Arabia South Africa Total Grand total 100% 100% 100% 100% 100% 100% 100% Note. The global market share of derivatives holdings for each dealer is calculated based on the dealer s notional amount outstanding of its reported total derivatives. A dash,, indicates no reporting dealers, and 0.00 indicates a number less than 0.005%. (Continued)

12 50 Emm and Gay TABLE II Geographic Composition of OTC Derivatives Dealers (Averaged over ) (Continued) Ranking By number of dealers By market share (%) Difference Rank Average Rank Average in ranking Country (1) (2) (3) (4) (1 3) Panel C: Comparison of country rankings US Germany Japan Britain France Italy Canada Australia Belgium Switzerland Denmark Austria Netherlands Spain South Africa Portugal Singapore Sweden Finland Bahrain Norway Ireland Saudi Arabia Poland Greece Israel Korea Thailand Brazil Czech Malaysia China Russia Hong Kong Note. This panel presents a rank order comparison of two measures: the average annual number of dealers and their average annual global market share based on statistics reported in Panels A and B, respectively. Both measures are calculated for each country over the period. A positive (negative) difference in rankings indicates the degree of improvement (lowering) of a country s ranking on the basis of market share relative to its ranking by number of dealers.

13 Global Market for OTC Derivatives 51 operations and subsidiaries of a number of U.S. and European-based dealers, such as HSBC (originally the Hong Kong and Shanghai Banking Corporation), Standard Chartered, JPMorganChase and Deutsche Bank (see, e.g., Farooqi, 2002). Other countries, which include Bahrain, Brazil, Saudi Arabia, and South Africa, have a small contingent of dealers. In Panel B of Table II we report for each country and year the cumulative notional holdings of all dealer positions expressed as a percentage of the global total. Because position information is reported in the database on a consolidated basis, we attribute the entirety of a dealer s positions to its home country, though we recognize that dealers may engage in significant cross-border activity. Loosely speaking, we see that dealer activity measured on the basis of market share appears somewhat consistent with that using simple number of dealer totals. One difference is that U.S. dealer activity becomes relatively larger. That is, the United States has only about one-fifth of the global number of dealers, but those dealers conduct more than one-third of the business. Using the year 2000 as a basis for reference, the United States is shown to have approximately 38% market share of the global total. Other leading countries include Germany (13%) followed by Japan (12%), France (11%), Britain (8%) and Switzerland (6%). Together, these six countries account for roughly 88% of the global total. By geographic region, in year 2000 Europe accounted for the largest market share with 46%, followed by North America with 40%, and the Asia/Pacific region with 13%. Looking at trends over our sample period, the market shares of North America and Europe have been somewhat stable with slight increases. The North American increase is driven by the United States and, within Europe, largely by Germany. Further, these increases coincide with a decline in the market share held by Japanese dealers, which has shown a continual drop from a peak of 21% in 1995 to a 9% share in Japan s 21% share in 1995 placed it second in the world, behind only the 28% share of U.S. dealers. In Panel C we provide additional perspective as to the relative number of dealers and global market share of each country by comparing each of the two ranking measures. Using data from Panels A and B, we first calculate the average annual number of dealers and the average annual market share for the period and then report the rank ordering based on each of these two measures. In the last column we compute the difference in rankings under the two measures. A positive (negative) difference in rankings indicates the degree of improvement

14 52 Emm and Gay (lowering) of a country s ranking on the basis of market share relative to its ranking by number of dealers. The same five countries rank in the top 5 under both measures, though Germany notably fell two spots under the market share measure relative to the number of dealers measure. For other countries ranked in the top 20, countries whose rankings by market share fell significantly below their rankings by numbers of dealers include Italy and Australia ( 5 each), Austria, Portugal and Singapore ( 4 each) and South Africa ( 3). Countries showing sizeable positive increases in relative rankings using market share include Sweden ( 8), Switzerland, Netherlands, Finland, and Norway ( 4 each). Global and U.S. Dealer Concentration Levels Using our global market share percentages computed at the individual dealer level, we next estimate for each year the N-dealer concentration ratios. We repeat these calculations after restricting our sample to U.S. dealers and U.S. market totals. These statistics for years 1995 and 2000 are presented graphically in Figure As we observe in Figure 1, for both samples the levels of concentrations have grown significantly over the years as indicated by the upward shift in the two sets of curves. To illustrate, for the global sample of dealers (see the GL 95 and GL 00 curves) in 1995 the four-firm ratio was 14%, the eight-firm ratio was 23%, and the 20-firm ratio was 48%. In 2000 these percentages had grown to 28%, 42%, and 67%, respectively. For U.S. dealers, concentration levels (see US 95 and US 00) are significantly higher than those measured on a global basis and have also increased over time. In 1995 the four-firm U.S. concentration ratio was 45%, the eight-firm ratio was 70%, and the 20-firm ratio was 98%. By 2000 these had grown to 69%, 89%, and 98%, respectively. In Table III we provide information regarding the leading dealers underlying each pair of concentration curves illustrated in Figure 1. In Panel A we list the 10 leading dealers on a global basis along with their relative market shares for years 1995 and Similarly, Panel B lists the 10 leading U.S. dealers in 1995 and 2000 and their U.S. market 12 The results for all other years are available upon request. Also, we present results for the year 2000 instead of 2001 because of the absence of holdings data in 2001 of two large dealers (Goldman Sachs and Morgan Stanley) as discussed earlier in footnote Smithson (1995, 1996) reports ranking information of leading OTC dealers for the earlier period of

15 Global Market for OTC Derivatives 53 Cumulative market share 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% N-dealer concentration ratio GL '95 GL '00 US '95 US '00 FIGURE 1 Global and U.S. dealer concentration ratios: 1995 versus Note. The figure shows N-dealer concentration ratios for the global and U.S. samples of OTC derivatives dealers, using corresponding market share percentages computed at the individual dealer level. The concentration ratios are presented for years 1995 and The global-market sample contains 217 dealers in 1995 and 199 in The U.S.-market sample contains 47 dealers in 1995 and 41 in All statistics are shown for the first 20-firm concentration ratios. GL 95 and GL 00 refer to the 1995 and 2000 global concentration ratios, respectively; while US 95 and US 00 refer to the 1995 and 2000 U.S. concentration ratios, respectively. shares. Comparing the 2000 rankings with those in 1995, we see large changes reflecting, in part, several incidences of industry consolidation. The most significant of these involved Chase Manhattan Bank. In 1995 Chase Manhattan had a global ranking of 21st and market share of 1.68%. Following its subsequent mergers and takeovers involving Chemical Bank, Robert Fleming (U.K.) and J.P. Morgan, JPMorganChase had become by the year 2000 the global leader with a 13.9% global market share and 36.9% U.S. market share. We explore dealer merger activity in greater detail in the next section. Although the concentration numbers just presented appear high, we present them solely as another indicator of industry structure. We acknowledge that a number of concerns have been expressed by policy makers and market observers with respect to the growth and size of the derivatives market and level of concentration therein. Many of these have centered on systemic risk concerns, that is, the risk that a default by a major dealer could cause a domino effect, affecting not only the well-being of immediate counterparties, but spreading and ultimately

16 54 Emm and Gay TABLE III Leading Global and U.S. Dealers: 1995 versus Market Market RankingDealer share (%) RankingDealer share (%) Panel A: Global dealer rankings based on global market share 1 Chemical (US) JP Morgan Chase (US) JP Morgan (US) Deutsche Bank (Germany) Societe Generale (France) Citigroup (US) Citicorp (US) Bank of America (US) Fuji Bank (Japan) BNP Paribas (France) Credit Suisse (Switzerland) Goldman Sachs (US) NatWest Bank (Britain) Royal Bank of Scotland Credit Lyonnais (France) 2.38 (Britain) 9 Swiss Bank Corporation Fuji Bank (Japan) 3.19 (Switzerland) 9 UBS (Switzerland) Industrial Bank of Japan Societe Generale (France) 2.88 (Japan) Panel B: U.S. dealer rankings based on U.S. market share 1 Chemical JP Morgan Chase JP Morgan Citigroup Citicorp Bank of America Bankers Trust Goldman Sachs Merrill Lynch Morgan Stanley BankAmerica Merrill Lynch Goldman Sachs Lehman Brothers Chase Manhattan American International Lehman Brothers 5.61 Group 10 Salomon Berkshire Hathaway 1.46 (General Re) 10 Bank One Corporation 1.29 threatening the entire financial system. 14 However, we note the number of safeguards in place to help prevent such occurrences including regulatory initiatives, such as bank examinations and capital adequacy standards. More important, the dealer community has been proactive in making important advances with the development and use of master agreements, bilateral netting agreements, collateral arrangements, and other risk-mitigation arrangements. For further discussion of these market-based mechanisms for addressing counterparty risk, see, for example, the Group of Thirty (1993), Gay and Medero (1996), and 14 See Hentschel and Smith (1995) for an analysis of why systemic risk concerns attributable to derivatives have been overstated.

17 Global Market for OTC Derivatives 55 Weinstein (2003). Also, Bomfim (2002) finds empirical evidence that netting agreements and other credit enhancement mechanisms used in swaps markets have been successful in mitigating counterparty credit risk during periods of market turmoil. MERGER ACTIVITY We next investigate the extent and implications of merger activity among dealers (see Appendix A for information on dealers that experienced mergers). During the time period, we identify a total of 54 mergers among dealers. A breakdown of the level of yearly merger activity is provided in the first row of Table IV. The years 1997 and 1999 had the highest occurrence of mergers with 14 and 15 mergers, respectively. We alternatively measure the magnitude of merger activity by computing separately the cumulative premerger market shares of both target and acquiring firms. That is, for each year, we sum the market shares of target firms and also that of the acquirers. These statistics are reported in the second and third rows, respectively, of Table IV. For target firms, with the exception of 1996 and 2001, merger activity in each year was comparable in magnitude with the cumulative premerger market share in the 5% 6% range. For acquirers, 2000 and 1998 were particularly active years, with acquirers having premerger market shares of 16.9% and 9.4%, respectively. The year 2000 results are driven largely by the completion on December 31, of the merger between JPMorgan and Chase TABLE IV Dealer Merger Activity: Number of mergers (54 in total) Cumulative premerger market share Target firms 6.66% 0.08% 4.75% 5.68% 6.52% 5.81% 0.02% Acquiring firms 4.64% 3.11% 6.71% 9.42% 7.93% 16.86% 0.49% Total 11.30% 3.19% 11.46% 15.10% 14.45% 22.67% 0.52% Note. The table presents the annual number of mergers and the corresponding cumulative pre-merger market shares of the acquiring and target firms. The sample is comprised of 54 dealer mergers during the period as reported in the Database of Dealer Outstandings published by Swaps Monitor Publications, Inc.

18 56 Emm and Gay Manhattan. This was by far the largest merger among derivatives dealers to date and served to solidify Chase s position as the largest derivatives dealer in the world. To illustrate, in 1999 Chase Manhattan held the top position with 8.1% of the global dealer market share, while JP Morgan ranked third with 5.4%. In addition to having the highest global market share based on all holdings, Chase also ranked first in interest rate derivatives, third in currency derivatives, twelfth in equity derivatives, fourth in commodity derivatives, and third in credit derivatives. Following the merger, JPMorganChase s global share rose to 13.9% and it ranked first in interest rate, currency, and credit derivatives; second in equity derivatives; and seventh in commodity derivatives. In Table V we report on the nature of the merger activity on a geographic basis (i.e., U.S., European, and other). We first note that most merger activity has occurred among European dealers. This entailed 35 of the 54 total mergers followed by nine mergers among U.S. dealers. Second, we observe only a few mergers that did not involve either U.S. or European firms (see other ), because U.S. and European dealers were involved in all but 3 of the 54 mergers. Third, the number of interregional mergers of dealers was relatively few. There were four instances of European acquirers taking over U.S. firms, but only two instances of a U.S. dealer acquiring a European target. The most significant case of the former was the Deutsche Bank acquisition of Bankers Trust in These two dealers had premarket shares of 3.4% and 1.7%, which corresponded to global rankings of fifth and eighteenth, respectively. TABLE V Geographic Breakdown of Dealer Mergers: Acquirer US European Other Total Target US European 2 35 a 1 b 38 Other c 3 Total Note. The table provides a locational breakdown of acquirers and targets involved in 54 dealer mergers over the period as reported in the Database of Dealer Outstandings published by Swaps Monitor Publications, Inc. a European mergers were intercountry and 26 were intracountry. Out of nine intercountry mergers, three mergers involved a non-european Union acquirer that was the same Finnish dealer. b This was a merger between Bahraini and British derivatives dealers. c Two mergers were between Japanese derivatives dealers and one was between Australian dealers.

19 Global Market for OTC Derivatives 57 We next investigate the extent to which merger activity had an effect on the subsequent product line offerings of the merged entity. For each merger we computed four statistics regarding the subsequent number of product line offerings of the merged entity measured one year (or immediately thereafter) following the merger. These include (1) the number of product lines that had been offered by the target that were subsequently dropped by the acquirer (lines dropped), (2) the number of product lines that had been offered by the acquirer and were exited (lines exited), (3) the number of new products lines offered by the merged entity that had not been formerly offered by the acquirer but were offered by the target (lines added), and (4) the number of new product lines that had not been formerly offered by either the target or acquirer (lines blossomed). We find that for product lines dropped, of the 54 mergers there was one line dropped in five cases, and two lines dropped in two cases. In the remaining 47 mergers, there were no lines dropped. For product lines exited, three acquirers exited one of their existing product lines, while one acquirer exited two product lines. For product lines added, there was one addition in four mergers, and two additions in two mergers. Finally, for product lines blossomed, in nine instances there was an increase of one new product line, and in one instance the merger was followed by the addition of two new product lines. Thus, in sum, it appears that neither the offering of new product lines nor savings from reducing product lines appear to be a primary motive for or a consequence of the majority of the mergers. EXTENT OF ASSET GROUP AND PRODUCT LINE PARTICIPATION In this section we further analyze the extent of dealer participation in each of the various (1) asset groupings, and (2) product lines. As mentioned earlier, Swaps Monitor reports breakdowns of dealer holdings in five asset groups: interest rate, currency, equity, commodity, and credit derivatives. Also when available, further breakdowns within each asset group (with the exception of credit derivatives) are also provided along three product lines: swaps, forwards, and options. Asset Group Analysis We compute and report in Table VI the average annual number of dealers reporting positions for each asset group over the period. This is done for both the global universe of dealers as well as for each

20 58 Emm and Gay TABLE VI Numbers of OTC Derivatives Dealers by Asset Group and Geographic Region (Averaged over ) North Global America Europe Asia/Pacific Other Interest rate derivatives Currency derivatives Equity derivatives Commodity derivatives Credit derivatives Total number of dealers Note. The table presents the average annual number of dealers reporting positions in each asset group over the period The numbers are reported for the global population of dealers and those in various geographic regions. geographic region, that is, North America, Europe, Asia/Pacific, and other. On the global level, the two asset groups having the greatest degree of dealer participation are interest rate and currency derivatives. These groups have comparable averages of and participating dealers. Equity derivatives run a distant third (116.6 dealers), followed by commodity derivatives (66.7 dealers), and then credit derivatives (27.9 dealers). These orderings are also generally observed within each of the geographic regions. One exception is that within North America, the number of dealers offering commodity derivatives exceeds that for equity. Although credit derivatives have the smallest number of dealers, they have shown rapid growth in recent years. Our yearly analysis shows that dealers offering credit derivatives had grown from zero in 1995 to 59 in 2001, attesting to the rise in popularity of these products. To explore further the nature of dealer participation in various asset group combinations, we address two related questions. First, what proportion of dealers makes markets in various combinations of derivative asset groups? For example, what fraction of all dealers makes markets in both interest rate and currency derivatives? Second, if we observe a dealer who makes a market in one asset group, what is the likelihood that the same dealer makes a market in another specified group? For example, given that a dealer offers commodity derivatives, what is the likelihood that the same dealer also offers equity derivatives? We perform these calculations on a yearly basis for all pairwise group combinations and report averages across years. The calculations are conducted

21 Global Market for OTC Derivatives 59 TABLE VII Derivatives Dealers Offerings: Pair-wise Product Proportions (Averaged over ) Asset group Interest rate Currency Equity Commodity Credit Panel A: Global Interest rate Currency Equity Commodity Credit Panel B: North America Interest rate Currency Equity Commodity Credit Panel C: Europe Interest rate Currency Equity Commodity Credit Panel D: Asia/Pacific Interest rate Currency Equity Commodity Credit Panel E: Other Interest rate Currency Equity Commodity Credit Note. The table presents the average fraction of dealers who make markets in various pair-wise derivative asset group combinations over the period. The panels are constructed using the global sample of dealers and those in each of the geographic regions. using the global population of dealers and separately for each geographic region. The results for both questions are presented in Tables VII and VIII, respectively. In Panel A of Table VII, for the global set of dealers, we see that 92.6% of all dealers offered both interest rate and currency derivatives,

22 60 Emm and Gay TABLE VIII Derivatives Dealers Offerings: Conditional Product Proportions (Averaged over ) Asset group North Conditioned Paired Global America Europe Asia/Pacific Other Interest rate Currency Equity Commodity Credit Currency Interest rate Equity Commodity Credit Equity Interest rate Currency Commodity Credit Commodity Interest rate Currency Equity Credit Credit Interest rate Currency Equity Commodity Note. The table reports the average fraction of dealers who, given that they make markets in one asset group, also make a market in a second asset group. These proportions are averaged over the period for the global sample of dealers and those in each geographic region. which is by far the highest reported pairwise combination. (Note that numbers along the diagonal simply represent the fraction of dealers offering derivatives in one specific asset group.) More than one-half of all dealers (55%) made markets in both interest rate and equity derivatives as well as in both currency and equity derivatives. Approximately 31% offered both commodity and interest rate derivatives as well as commodity and currency derivatives, whereas only 25% of dealers offered both equity and commodity products. Across each of the geographic regions (Panels B E), we see a fairly high percentage of dealers offering both interest rate and currency derivatives. These products appear to be staple offerings of most dealers anywhere. Comparable fractions of dealers (more than 50%) in