corporate bond e-trading - mckinsey/media/mckinsey/dotcom/client... · 2020. 8. 5. · 2 corporate...

Corporate Bond E-Trading: Same Game, New Playing Field

Contents

Executive Summary 2

Introduction 4

Liquidity: Not So Bad, but Challenges Lurk 7

Corporate Bonds and Order-Driven 10E-Trading: Not a Match

The Electronic Road Ahead 14

Equities-Like Trading: What It Would Take 17

Winners and Also-Rans 19

Preparing for the E-Trading Future 20

Corporate Bond E-Trading: Same Game, New Playing Field

2 Corporate Bond E-Trading: Same Game, New Playing Field

With a tumultuous mid-year 2013 correctionin the bond markets heightening anxietyamong corporate bond market participants,the industry continues to hold out hope thatelectronic trading (e-trading) will be thepanacea for the market’s ills. However, newresearch from McKinsey & Company andGreenwich Associates suggests that full-fledged e-trading in corporate bond marketswill be slow to arrive (if at all). Despite thelaunch of several corporate bond tradingplatforms since spring 2012, the market isunlikely to ever resemble cash equities oreven foreign exchange. A spring 2013 surveyof 117 institutional corporate bond investorsin the United States and Europe, under-pinned by in-depth interviews with assetmanagers, leading banks and market opera-tors, reveals that:

• Until mid-spring 2013, the liquidity short-age, a spur for many of the new e-tradingplatforms, was not as bad as many hadfeared. Around 30 percent of survey re-spondents said that liquidity had actuallyimproved during the prior 18 months. Butmost respondents also felt that liquiditywas not entirely healthy, as a rise in

issuance obscured some underlying prob-lems. With the phase-in of post-crisis regu-lation, a clear majority expect liquidity todeteriorate further.

• Few participants foresee a revolution ine-trading. When asked about the next fiveyears, 80 percent said the multi-dealer re-quest-for-quote (RFQ) platforms onwhich most e-trading has been conductedfor years will continue to prevail.

• Buy-siders are generally reserved aboutother e-market models. A surprisinglyhigh number, 25 percent, expect crossingsystems to dominate in five years. Not un-expectedly, the figure for single-dealerplatforms is lower, at 10 percent, whilefewer than 10 percent were enthusiasticabout the prospects for exchange-oper-ated platforms.

The skepticism we observed regarding cor-porate bond e-trading is a recognition ofboth the structural differences between thecorporate bond market (historically quote-driven) and e-trading pioneer markets likecash equities (largely order-driven), and theprofound changes in industry conduct that

Executive Summary

3Corporate Bond E-Trading: Same Game, New Playing Field

would be required to make bona fide match-based e-trading come to life.

Furthermore, the liquidity profile of thecorporate bond market and the scarcity ofdesktop “real estate” on the buy side implythat the winning e-market models will ac-commodate only a few centralized plat-forms, as a proliferation of platforms

would likely result in harmful liquidityfragmentation.

Corporate bond e-trading may be off to aslow start, but a gradual transformation inhow the market trades is underway, andmarket participants will need to adapt.Dealers must take steps—if they have notalready—to, among other things, align theorganization, harness the opportunity inbetter data management, transform the salesforce and reduce cost per trade. Asset man-agers should rethink their investment deci-sion-making process and revamp the tradingfunction. And market operators must workto enhance their platforms to deliver moreof the services that corporate bond dealersand buy-siders seek.

Despite the launch of several corporate bond trading platforms since spring 2012, the market is

unlikely to ever resemble cash equitiesor even foreign exchange.


Although anticipated by some, the mid-year2013 sell-off that upended the bond marketsnonetheless proved stunning in its ferocity.While some semblance of calm appears tohave returned to the markets, corporate bondmarket participants remain concerned, as thatmarket was already dealing with a raft ofchallenges prior to the sell-off. Regulatory re-form had handcuffed dealers, some say, byraising capital requirements; dealers re-sponded by cutting net inventory. The conver-gence of buy-side investment strategies and theproliferation of buy-and-hold investors culti-vated a “one way” market, in which it was in-creasingly difficult to find sellers. As expected,that pattern inverted during the sell-off: in-stead of offers, bids became harder to find—though the massive repricing has begun toattract renewed buying interest from yield-hungry investors.

It looks like some big bills are coming due—even without taking the recent bond marketcorrection into account. Historically low in-terest rates in recent years fueled an issuanceboom, which was good for the debt capitalmarkets (DCM) business and for “on therun” trading. But that same issuance boom,according to the Securities Industry and Fi-nancial Markets Association (SIFMA), alsodrove outstanding U.S. corporate debt to $9trillion by the end of 2012, a record high,

and 2.2 times higher than it was in 2002.There is a tremendous and potentially unsus-tainable amount of paper in investors’ hands,and this harsh reality is causing much angst.

Anticipating the market’s next phase is vitalto dealers. According to Coalition, sales &trading in cash credit accounted for $4.5 bil-lion of the $17 billion generated by the 10largest investment banks in overall creditsales & trading in 2012. Furthermore, inter-dependencies with other businesses, such asDCM, foreign exchange and rates, make itimperative for dealers to remain relevant incash credit trading—cost challenges aside.

By way of proof, several market participantshave announced or launched new e-tradingplatforms and systems. BlackRock announcedplans in April 2012 to launch a crossing sys-tem called Aladdin Trading Network. In April2013, it announced a new plan for the sys-tem, under which it would team up withMarketAxess Holdings, itself the operator ofa corporate bond e-trading platform. In June2012, Goldman Sachs brought GSessions, an-other crossing system, to market. DeutscheBank has introduced the concept of a liquidityhub. Several other similar solutions are in theworks. Many in the industry see these solu-tions as a potential antidote for the industry’sstructural problems, and—now perhaps—even as a way to calm roiling markets.

Introduction


Will the new e-trading platforms right theship? Or are other steps needed? To find out,McKinsey and Greenwich Associates collab-orated on a research effort on the future ofcorporate bond e-trading. In the spring of2013, we conducted an online survey of 117buy-side portfolio managers, traders and an-alysts (35 from the United States and 82from Europe). We also conducted in-persondiscussions with leading buy- and sell-sideparticipants, including 8 of the 10 largestdealers, as well as the operators of severalmajor e-trading platforms.

This research revealed that true corporatebond e-trading is a long way from becoming

reality, with nascent “e” activity thus farpredominantly dealer-driven. Other assetclasses are further advanced (Exhibit 1).While e-trading will undoubtedly play animportant role in the future of corporatebond trading, structural realities stand in theway of attaining the order-driven “nirvana”that the cash equities market has achieved.Corporate bond markets in the aggregateare unsuited to e-trading, and the marketparticipants surveyed and interviewed aredubious about the prospects for bona fidematch-based e-trading.

To be clear, no tweaks to market structure canslow down the massive repricing of corporate

E-p

en

etr

ati

on

to

da

y

Current state of evolution

Bespoke productProduct High product variability Products becoming standardized

Highly standardized; “vanilla”

Complex pricingPricing Complex pricing Basic pricing Transparent, clear pricing

IlliquidLiquidity Somewhat liquid Liquid Extremely liquid

Limited SDPs4/MDPs5 Platform Basic SDPs/MDPs Mature SDPs/MDPs Streaming largely available

Strong balance sheet, credit rating and risk management

Success drivers



Low cost per trade and large volumes, given low margins

Treasury futures (up to 90%)Cash equities (70-80%)iTraxx CDS index (70%)Foreign exchange, spot (55–65%)

European government bonds (40–50%)Foreign exchange, forwards (40%)Covered bonds (40%)Precious metals (40%)

U.S. Treasur-ies (30%)Short-term interest rate trading (30%)Foreign exchange, swaps (30%)

Bespoke interest rate swapsStructured credit/rates

IG1 cash (15–20%)HY2 cash (10%)

CDS,3 single name

Standard-ized interest rate swaps (20–25%)Repos (25%)

Foreign exchange, options (15-20%)

Largely voice Moving toward “e” Past inflection point

Fully electronic/“futurized”

Level of penetration by volume

1 Investment grade corporate bonds 2 High yield corporate bonds 3 Credit default swaps 4 Single-dealer platforms 5 Multi-dealer platforms Source: McKinsey & Company

Exhibit 1

Corporate bond trading is still in the nascent stages of electronification


bond portfolios ordained by the market fromtime to time, like the one fueled by FederalReserve Chairman Ben Bernanke’s May 22ndand June 19th (2013) remarks that the Fedcould begin tapering its bond purchases laterin the year. Barring a drastic reduction in cap-ital requirements for corporate bonds, the sellside will be severely limited in its ability tobackstop the market; the pendulum may wellswing back in the other direction.

* * *

This paper first examines the recent liquiditydevelopments in the corporate bond market,where several underlying problems may nowbe coming to a head. It then outlines thestructural barriers that stand in the way offurther electronification. Following this is adiscussion of industry views on several e-trad-ing models, including those that have recentlydebuted. Most respondents to the McKinsey-Greenwich Associates survey expect that

multi-dealer RFQ platforms will continue todominate e-trading for some time.

The paper then outlines—under the assump-tion, however tenuous, that structural barri-ers can be overcome—the dramatic changeson the part of market participants thatwould be required for match-based corpo-rate bond e-trading to come to life in theway it has in cash equities and a few otherasset classes like U.S. Treasuries, CDS indicesand foreign exchange. Absent those changes,we offer a view of how the market’s dynam-ics are likely to favor certain participants, es-pecially independent operators ofquote-driven markets. Finally, the paper pro-poses a number of ideas for how dealers,asset managers and independent market op-erators can capitalize on the current situa-tion, and lay the groundwork for the futureof corporate bond e-trading. Throughout, thepaper focuses primarily on the U.S. market.


For all the long-standing worries about liq-uidity, it actually held up rather well throughmid-spring 2013. In both the United Statesand Europe, in both investment grade (IG)and high yield (HY) corporate bonds, mostsurvey respondents said they had not seensignificant deterioration in the quality of liq-uidity—ease of execution, price impact of atrade, width of the bid-ask spread, and soon—over the previous 18 months. In fact,about 30 percent of respondents said theyhad observed some improvement in IG andHY corporate bond liquidity over that pe-riod. To be fair, 45 percent noted degrada-tion, but only 10 percent deemed thedeterioration significant.

Looked at another way, however, liquidity in-deed weakened. In the United States, for ex-ample, IG and HY corporate bond turnover(the value of bonds traded divided by theamount outstanding) in the aggregate failedto keep pace with the issuance boom, fallingaround 20 percentage points from pre-crisis2007 to around 85 percent in 2012.

However, with outstandings up 50 percentand net corporate securities inventory held byprimary dealers—often informally referred toas corporate bond inventory—down around80 percent since 2007, one might have ex-pected U.S. IG and HY corporate bond

turnover to have slowed much more than itdid (Exhibit 2, page 8). It appears that the ex-pected impact on liquidity of the drop-off innet inventory was modest and was certainlyless severe than some of the dire predictions.

To be sure, some of the moves initiated bydealers to cut net inventory hurt liquidity. Aregulation-induced wind-down of proprietarytrading, exits by some bulge bracket dealersand a decline in basis trading of credit defaultswaps (CDS) and corporate bonds all madethe market less liquid. But other inventory-reducing measures actually buttressed liquid-ity. As a case in point, for the remainingbulge bracket dealers, the velocity of their in-ventory turnover actually increased, allowingthem to trade more with less inventory.

We would be remiss not to point out that thepre-crisis rise and post-crisis “fall fromgrace” of non-agency mortgage-related secu-rities—which the Federal Reserve included inits definition of net corporate securities in-ventory until April of this year—played arole in the corresponding spike and drop innet inventory. It is safe to say that thischange in net inventory did not have a directimpact on corporate bond liquidity.

Another sign of the liquidity challenges un-derlying the corporate bond market can be

Liquidity: Not So Bad, But Challenges Lurk


found in the significant share of trading in2012 that was done in recent issues. In fact,according to BlackRock Investment Institute,highly liquid corporate bonds carried a pre-mium (defined as the spread between liquidand less liquid U.S. IG corporate bonds ofthe same issuer with similar maturity) ofabout 16 basis points over their less liquidcounterparts in mid-2012. That premiumwas about 10 basis points higher than theaverage of the preceding seven years—a pre-mium that likely would have been evenhigher in the absence of a low-volatility envi-ronment, driven in part by the Federal Re-serve’s quantitative easing campaign.

Is the tide turning?

In sum, liquidity fared better than expecteddespite the stresses that imperiled its founda-tions—MarketAxess bid-ask spreads in U.S.IG corporate bonds through mid-spring

2013 were not dramatically higher than theywere prior to the global financial crisis (Ex-hibit 3). During the recent sell-off, however,bid-ask spreads widened—not a huge sur-prise given the magnitude of the bond mar-ket correction.

Although the jury’s still out on how the mid-year tumult in the corporate bond marketwill ultimately play out, the liquidity skepticsmay have the last word. The issuance boomand follow-on trading in these issues havelikely buffered the market against a steeperdrop in turnover rates. Consequently, the re-cent slowdown in new issuance in the UnitedStates—June 2013 levels fell almost 30 per-cent relative to June 20121—may not augurwell for corporate bond liquidity. As finan-cial market reforms continue to be phased in,liquidity could evaporate. Eighty percent ofU.S. respondents to our survey took thisdownbeat view of liquidity, as did 55 percent

~85%

Net corporate securities inventory1 and outstanding U.S. corporate debt, 2001–13

12-month rolling turnover of TRACE-eligible U.S. investment

grade and high yield corporate bonds 1 Comprises U.S. dollar-denominated debt securities issued by corporations incorporated in the U.S., including bonds, notes and debentures;

commercial paper; covered bonds; privately placed securities (e.g., 144A securities); non-agency collateralized mortgage obligations and real estate mortgage investment conduits; and non-agency stripped securities. Data through March 27, 2013.

Source: Federal Reserve Bank of New York; Securities Industry and Financial Markets Association; MarketAxess Research

0

2

4

6

8

1

3

5

7

9

10

0

50

100

150

200

250

Net corporate securities inventory held by primary dealers$ billion (left axis)

6/23/10 6/22/116/24/096/25/086/28/066/29/056/30/047/2/037/3/027/4/01

~105%

6/20/126/27/07

Outstanding U.S. corporate debt$ trillion (right axis)

Exhibit 2

Outstandings rose, net dealer inventory fell, but turnover dropped far less dramatically

1 Based on corporate debt issuancedata from the Securities Industry andFinancial Markets Association


of European respondents, who may be lesssensitive to the issue, having already experi-enced the pain of the sovereign debt crisis.

Among respondents who believe liquidityeroded during the 18 months preceding mid-spring 2013, over 80 percent cited post-crisissell-side developments as drivers of this dete-rioration. Many noted a decline in the ap-petite of bulge bracket dealers to facilitatetrading—primarily as a function of the risingprice of risk-weighted assets. According tosome of the largest buy-siders, it has becomeincreasingly difficult to get quotes on blocktrades,2 forcing them to break their blocksinto smaller trades and take this businesselectronic where possible. Another factorcited was a drop-off in proprietary trading,which has hurt liquidity in “off the run” is-sues. Finally, market exits by some bulgebracket dealers have further reduced balancesheet capacity, and perhaps even competition.

The profile of the buy side has also givenrise to concern. Until the recent sell-off, anincreasing share of investors had beenadopting a buy-and-hold stance, and in-vestor strategies were converging as otherstrategies exited the market. During the sell-off, many of these same investors weredumping (or trying to dump) their holdings.In both cases, investors moved in lockstep,creating a liquidity-challenged one-way mar-ket. As a matter of fact, about 80 percent ofEuropean survey participants cited buy-sideinvestment behavior as an underlying causeof worsening liquidity. However, only about50 percent of U.S. respondents shared thisview. Instead they singled out sell-side limi-tations as a driver of deteriorating liquidity.To some extent, the buy side’s greater de-pendence on bulge bracket dealers in theUnited States relative to Europe may explainthis divergence.

2 A block trade is a trade with a valuegreater than $5 million.

2003 2005 2007 2009 2011 2013 0

10

15

20

25

5

MarketAxess U.S. investment grade corporate bond bid-ask spreads, 2002–131

Bps

1 Based on investment grade corporate bonds that traded on both sides of the market on a given day on MarketAxess; a statistical process called a LOWESS function was applied to the data to create a line representing a smoothed-out average.

Source: MarketAxess Research

Exhibit 3

Liquidity fared reasonably well through mid-spring 2013, but bid-ask spreads widened during the recent sell-off


Two factors suggest that corporate bondtrading is poised to become more electronic(Exhibit 4). First, post-crisis trade sizes areshrinking, a trend best seen in the decline inblock trading. Dealers have less balancesheet capacity to take on these trades. As aresult, trading frequency is on the rise—thesecond factor. Interestingly, both trade sizeand frequency appear to have stabilized at anew steady state over the past four years.

Many market observers see these trends asharbingers of an increase in electronification.Of course, some corporate bond trading isalready electronic, in a sense. A 2012 Green-wich Associates survey of 550 institutionalinvestment professionals found that in theUnited States, e-trading as a percentage of alltrading volume in U.S. IG corporate bondsgrew from 10 percent in 2011 to 14 percentin 2012. In Europe, e-trading is more preva-lent, and expanded from 22 to 29 percent inthe same period, primarily because e-tradingeases cross-border activity. By far most ofthis e-trading takes place on multi-dealerRFQ platforms (Exhibit 5, page 12).

Not coincidentally, multi-dealer RFQ plat-form operators captured the lion’s share ofthe spike in U.S. corporate bond e-tradingvolumes observed during the recent sell-off.

According to The Wall Street Journal,Bloomberg’s corporate bond trading volumereached a record high for the company for themonth of June, while MarketAxess Holdings’smarket share of overall U.S. IG corporatebond trading hit 16.6 percent, also a recordfor the company for the month of June. Theimperative for highly motivated institutionalinvestors to access a broader swath of dealersin an increasingly capital-constrained environ-ment likely drove this development.

But true electronic match-based trading, likethat seen in cash equities, remains a distantdream. Why?

To begin, the asset class is much more het-erogeneous than equities. At its peak in1997, the U.S. stock market boasted about8,800 listed companies. That pales in com-parison to the U.S. corporate bond market,which had 37,000 publicly traded, TRACE3-eligible bonds outstanding in 2012. Thesheer number of issues greatly reduces theprobability of multilateral trade matching.

The structural differences between cash equi-ties and corporate bonds are manifest in thetwo assets’ very different trading profiles.Whereas the average U.S. stock tradedaround 3,800 times per day in 2012, the 13most liquid U.S. IG and 20 most liquid HY

Corporate Bonds and Order-Driven E-Trading: Not a Match

3 Financial Industry RegulatoryAuthority’s Trade Reporting andCompliance Engine (TRACE)


corporate bond issues traded only about 85times and 65 times per day on average, re-spectively (Exhibit 6, page 13). There are, infact, thousands of corporate bond issues thatrarely or never trade. Some facts:

• In 2012, 38 percent of the 37,000TRACE-eligible issues did not trade evenonce, with another 23 percent tradingonly a handful of times, as compared tothe 1 percent that traded every day, ac-cording to MarketAxess Research.

• In 2011, again according to MarketAxessResearch, 30 percent of the 32,000TRACE-eligible issues did not trade once,with another 26 percent trading only ahandful of times. Two percent tradedevery day.

• Over a period of about 1,150 trading daysbetween July 2002 and January 2007, 18percent of the more than 47,000 TRACE-eligible issues did not see any action, ac-cording to research by Michael Goldsteinof Babson College and Edith Hotchkiss ofBoston College.

Furthermore, when they do trade, corporatebonds involve a great deal more money thanthe average stock trade. The average tradesize for the most liquid U.S. IG securities isabout 70 times that of the average U.S. stocktrade. And according to McKinsey’s CapitalMarkets Trade Processing Survey, post-tradeprocessing costs in corporate bonds dwarfthose in cash equities. For e-trading to takeoff, trade sizes will likely need to be smaller.

Average trade size (par value traded)$ thousands

Q1 2010

Q3 2010

Q3 20

08

Q3 2011

Q1 2012

Q1 20

09

Q3 20

09

Q3 2012

Q1 20

07

Q1 2011

Q1 20

08

Q3 20

07

Average trading frequencyNumber of trades per day

Q3 2011

Q3 2012

50,000

45,000

40,000

35,000

30,000

25,000

20,000

15,000

10,000

5,000

0

1,000

900

800

700

600

500

400

300

200

100

0

Q1 2012

Q1 20

07

Q3 20

09

Q1 2010

Q1 2011

Q3 2010

Q1 20

08

Q3 20

07

Q1 20

09

Q3 20

08

Trading in investment grade and high yield U.S. corporate bonds, 2007–12

Source: Financial Industry Regulatory Authority’s Trade Reporting and Compliance Engine

Exhibit 4

Reduction in sell-side capacity for “balance sheet” trades is driving down trade size and boosting frequency


But their higher inherent costs make it ex-pensive to break down big corporate bondtrades into smaller ones.

It is also worth noting that corporate bondsare not as amenable to shorting as cash equi-ties. Consequently, it should come as no sur-prise that the corporate bond market is notas well suited for immediate two-way trad-ing—a key requirement for match-based e-trading—as its cash equities counterpart.

Finally, the attributes that the buy side val-ues most are better delivered through cur-rent favored channels than throughmatch-based e-trading. Buy-side respon-dents place the highest premium on imme-diacy (about 80 percent in the United

States and 85 percent in Europe) andanonymity (about 75 percent in the UnitedStates and 50 percent in Europe). As ex-pected, a good portion of the buy sidewould welcome streaming executable pric-ing (about 50 percent in both the UnitedStates and Europe). But they appear to likethe idea more in the abstract, when some-one else is providing the streaming pricing,than in reality. Given the press and excite-ment around initiatives such as Black-Rock’s Aladdin Trading Network,Goldman Sachs’s GSessions and UBS’sPrice Improvement Network (PIN), onemight have expected more than 30 percentof respondents in both the U.S. and Europeto value multilateral trading. Tellingly, no

United States Europe

90

32 4

1

36

17

72

1

Multi-dealer request-for-quote (RFQ) platform

Multi-dealer “odd lot” platform with live executable quotes

Single-dealer platform: eitherRFQ or with liveexecutable quotes

Central limit-order book

Crossing system

Question: “What share of your corporate bond trading volume do you currently trade through the following channels?”Percent, 20131

1 N = 109 total, 34 in U.S. and 75 in Europe. Figures may not sum to 100% because of rounding. Averages not weighted by respondents’ trading volumes.

Source: McKinsey-Greenwich Associates 2013 survey of institutional investors

Exhibit 5

Multi-dealer request-for-quote e-platforms dominate current corporate bond e-trading


one surveyed in the United States, and just2 percent of Europeans, views the suitabil-ity of an e-platform for algorithmic tradingas a critical attribute—a far cry from theworld of cash equities, where algorithmictrading has become the norm.

At the end of the day, the corporate bondmarket is—and could very well remain—aquote-driven (that is, dealer-driven) one, instark contrast to the largely order-drivencash equities market.

U.S. cash equities and U.S. corporate bonds, 2012

Average trading frequency per issueNumber of trades per day

Average value per trade$

~7,000

-99%

~400,000

~500,000

~85

-98%

~65

~3,800

NYSE and NASDAQ

stocks

Most liquid HY2 corporate

bonds

Most liquid IG1 corporate

bonds3

NYSE and NASDAQ

stocks

Most liquid HY corporate

bonds

Most liquid IG corporate

bonds

1 Investment grade 2 High yield 3 Defined as corporate bonds listed among both the top 50 issues in terms of trading frequency and in terms of value traded; 13 issues for IG, 20

issues for HY. Assumes 230 trading days in 2012. Source: NASDAQ OMX Group; NYSE Euronext; Financial Industry Regulatory Authority’s Trade Reporting and Compliance Engine

Exhibit 6

Even the most liquid corporate bonds are far less suitable for match-based e-trading than the average U.S. stock


Buy-siders are cautiously optimistic aboutcorporate bond e-trading. U.S. respondentsbelieve that 40 percent of corporate bondtrading volume can be executed electroni-cally; European respondents take a morebullish view, coming in at 65 percent. By2015, U.S. respondents expect to trade 30percent of their volume electronically andEuropeans expect to approach the limits ofelectronification, projecting a potentially un-realistic 60 percent—foreign exchange e-trading today tops out at 65 percent.

However, buy-siders foresee an evolution,rather than a revolution, and expect the mar-ket to remain predominantly dealer-driven inthe foreseeable future, albeit with multiple e-trading models to meet their needs. And theevolution will be narrow in scope. With themost liquid two deciles of outstanding U.S.corporate bond issues accounting for over 90percent of overall dollar-trading volume, cor-porate bond e-trading is unlikely to featurethe very long tail of issues that actually con-stitute the liquidity challenges.

In our survey, we asked respondents whichof the five e-channels they expect to predom-inate within the next five years: multi-dealerRFQ platforms, crossing systems, single-dealer platforms, multi-dealer odd-lot4 plat-forms or central limit-order books.

The buy side expects the status quo, with 70percent of U.S. respondents and 85 percent

of Europeans saying multi-dealer RFQ plat-forms such as those operated by Bloombergand MarketAxess Holdings will continue toprevail. These systems provide investors withwhat they need in the post-crisis era—theability to: trade efficiently at somewhatsmaller size with a broad pool of dealers; ex-ecute fairly quickly; fulfill “best execution”obligations (although potentially withgreater information leakage relative to tradi-tional voice- or messaging-based trading);and move toward straight-through process-ing. For dealers too, these systems work well.With balance sheet constraints limiting blocktrading, multi-dealer RFQ platforms are amore economical channel for transactingwith the long tail of small buy-side clients,trading in and out of positions, and process-ing smaller trades.

Market participants identified—albeit indi-rectly—the potential to improve upon themulti-dealer RFQ platform. More than 40percent of the bigger traders on the buy sideare open to the idea of providing quotes toRFQ platforms in an all-to-all environment,while some major dealers expressed some in-terest in receiving quotes from the buy side.In light of these developments, a further ex-pansion of complementary all-to-all func-tionality on multi-dealer RFQ platforms—something a few operators have alreadybegun experimenting with—could unlock ad-ditional liquidity. Armed with such function-

The Electronic Road Ahead

4 An odd-lot trade is a trade with a value ranging from $100,000 to $1 million.


ality, both buy- and sell-side participants cananonymously broadcast RFQs, with auto-mated alerts sent to participants with an in-terest in providing anonymous quotes for theunderlying issues. Of course, a cadre of clear-ing intermediaries would be needed to main-tain post-trade anonymity.

However, one obstacle to the successfuladoption of this all-to-all functionality is thebuy side’s limited willingness to disclose itstrading inventory to other buy-siders (evenanonymously); only one-quarter of respon-dents on both sides of the Atlantic expresseda willingness to do so.

Surprisingly, almost one-quarter of respon-dents (and 45 percent of the bigger traders)believe that crossing systems such as AladdinTrading Network, GSessions and PIN willplay a meaningful role in the future. Al-though dark pool operators in other assetclasses have disdained corporate bond trad-ing in the past, given the low likelihood ofnatural matching, over 50 percent of the big-ger traders still believe that a multilateralplatform open exclusively to the buy side cansucceed. Ultimately, the absence of a broadlyaccepted price discovery mechanism thatdoes not emanate from dealers (notwith-standing the efforts of independent vendors)

and the lack of a robust two-way market arethe twin elephants in the room.

With buy side-only crossing systems seem-ingly dead on arrival, what tweaks can bemade to give this e-market model—one withthe potential to limit information leakageand, hence, support trading in larger round-lot5 and perhaps even block sizes—a fightingchance? First, the platform must enlist sell-side “specialists” (as the cash equities mar-ket has) to support price discovery andpotentially sop up some of the buy or selloverhangs in assigned issues. Second, suchsystems should employ a call auction ratherthan a continuous-crossing format. Auctionsat designated times of the day will engenderbursts of liquidity, albeit primarily in themore actively traded corporate bonds, espe-cially around liquidity-driving events (forexample, earnings announcements and up-grades or downgrades in an issuer’s creditrating). Within this construct, the specialistsproviding the winning quotes in a competi-tive bidding process (i.e., the highest bid andlowest offer) for a specific auction could becompensated through trade value-basedmarkups or markdowns, with all tradescrossed at the midpoint of the bid-askspread. The trade-off for the liquidity im-provement is that call auctions would re-quire the buy side to sacrifice someimmediacy.

Dealers in such systems run a risk from pro-moting greater price transparency, whichmight allow the buy side to “free ride” andtrade amongst themselves at the midpoint ofthe bid-ask spread. To protect them, the plat-form would need to take steps, perhaps seek-ing binding commitments from buy-siders torefrain from such activity.

The buy side expects the status quo, with 70 percent of U.S.respondents and 85 percent of

Europeans saying multi-dealer RFQplatforms such as those operated byBloomberg and MarketAxess Holdings

will continue to prevail.

5 A round-lot trade is a trade with a value ranging from $1 million to $5 million.


As for platform governance, almost no onethinks a single dealer is the best owner/oper-ator of a crossing system; the perception ofconflicts of interest may account for the buyside’s resistance. Nor does leadership by aconsortium of buy-side players improve theiroutlook; no one surveyed in the UnitedStates, and just 5 percent of Europeans, likesthis model. Instead, 60 percent of the buyside in the United States and 50 percent inEurope prefer a centralized platform run byan independent operator, either as a singleowner or in partnership with a consortiumof market participants, perhaps because theybelieve that such a system minimizes liquid-ity fragmentation.

Less than 15 percent of U.S. respondents andless than 10 percent of European respon-dents think that single-dealer platforms (ei-ther RFQ or with live executable quotes)such as Barclays Automated Realtime Execu-tion (BARX) can become significant fixtures.The buy side’s desire for best execution andthe inefficiency of this platform for price dis-covery weigh heavily against the viability of

this market model. Furthermore, the capital-light post-crisis environment makes suchplatforms difficult to sustain; most banks areopting instead to focus their e-efforts onmulti-dealer platforms or, in a few cases,proprietary crossing systems.

Just 10 percent of traders in either region seea dominant future for multi-dealer odd-lotplatforms with live executable quotes, suchas BondDesk. But the institutional focus ofthe survey may partially explain the low tallyfor this historically retail-focused platform.Dealers may not have the appetite to providefirm quotes to informed investors, who fortheir part may not be interested in smaller-lot trading.

Critically for those in the industry who hopefor a rapid shift to bona fide e-trading, veryfew respondents (3 percent in the UnitedStates and 10 percent in Europe) were enthu-siastic about the prospects for central limit-order book systems such as NYSE Bonds andXetra Bonds. Exchanges have tried manytimes to launch or revive such platforms, butsuccess has proven elusive. Fewer than 15percent of respondents say they are willingto provide firm quotes on such systems,making it difficult to attract liquidity.

Irrespective of which e-market models ulti-mately win, the corporate bond market is notliquid enough and the buy side does not haveenough desktop real estate or the infrastruc-ture investment appetite to support morethan a couple of platforms—centralized plat-forms, that is—for each winning model.

Critically for those in the industry who hope for a rapid shift to bona fide

e-trading, very few respondents were enthusiastic about the prospectsfor central limit-order book systems

such as NYSE Bonds and Xetra Bonds.


In light of structural barriers, it is currentlydifficult to imagine a path forward for equities-style corporate bond e-trading. How-ever, if there were such a path, it would be anarduous one requiring behavioral and strategicchanges from all market participants: dealers,institutional investors, issuers and regulators.

Dealers

The support of dealers (especially bulgebracket dealers) would be essential. Interest-ingly, about 60 percent of U.S. respondentsexpect the smaller dealers, those with re-gional and local franchises, to step up and fillthe liquidity gap (versus only about 30 per-cent in Europe, where several national cham-pions already play a more important role).However, the top 10 dealers accounted forthe same share of overall dollar-trading vol-ume in U.S. corporate bonds in 2012 as theydid in 2009 (52 percent), demonstrating thatany match-based solution will require theprice discovery and quoting support of thebulge bracket. As such, the bulge bracketdealers would need to assume the role of des-ignated specialists or market makers on mul-tilateral platforms, and find ways to lowerclearing and settlement costs to makesmaller-lot trading economically feasible.However, they will have little incentive to

support unbridled, spread-compressing pricetransparency, unless one of the following sce-narios were to occur:

• Smaller dealers or other players (such asindependent “prop shops” and hedgefunds) seriously threaten to displace thetop tier. That movement may have alreadybegun: according to Bloomberg, credit-focused hedge funds have attracted morethan $100 billion in recent years and hiredtop traders away from the sell side. Theseless-regulated players are now in a posi-tion to compete with the big banks—readyand willing to provide intraday liquidity.But to succeed, they will have to overcomethe bulge bracket’s advantage in DCM andits experience in transacting with the buyside, which translate into superior issueknowledge and sourcing know-how.

• Bulge bracket dealers “get religion” andbegin to believe that volumes will rise ex-ponentially, thereby offsetting the antici-pated spread compression associated withmatch-based trading.

Even with the support of bulge bracket deal-ers, smaller dealers would need to providesupplemental liquidity, taking advantage oftheir lower capital requirements to carrymore inventory.

Equities-Like Trading: What It Would Take


Institutional investors

Institutional investors would have to collec-tively change their behavior in several ways.They would have to provide meaningful liq-uidity—for example, by placing firm quotesin central limit-order books. They wouldhave to accommodate “blotter scraping,” inwhich they anonymously share their buy andsell requests to increase the probability offinding a counterparty. Investors would needto trade in smaller lots, which would in-crease trade frequency. Investors in the ag-gregate would need to diversify their pool ofinvestment strategies to support a two-waymarket. Finally, they would have to showmore discipline in backing new issues; ulti-mately, investors are the consumers of theseproducts and are in the strongest position toinfluence issuer behavior.

Issuers

Issuers would need to standardize their prod-ucts. Match-based e-trading requires a highdegree of standardization to reduce issuefragmentation. But it will likely take a signifi-cant liquidity-damaging event (such as a pan-icked and protracted sell-off in the corporatebond market) or a concerted effort by thebuy side to spurn non-standardized issues, ei-ther of which would raise issuers’ effectivecost of funding, for companies to begin stan-dardizing their issues. Even then, 60 percent ofU.S. respondents and 45 percent of European

respondents are not convinced that a rise inliquidity premiums would convince compa-nies to standardize their issues.

Regulators

Some industry insiders believe that regula-tion will ultimately break the bottleneck, aswas the case for cash equities, where Regula-tion National Market System (Reg NMS),Regulation Alternative Trading Systems (RegATS) and decimalization spurred e-trading.However, the cash equities market was al-ready well suited to match-based e-trading,and the regulation simply removed theshackles from a market ready for change. Asdiscussed earlier, corporate bond trading isnot even close to being similarly ready. Andeven in cash equities, the uniform approachto large- and small-cap stocks may have hurtliquidity in the smaller securities, as frag-mentation and insufficient market maker in-centives continue to take a toll.

But some change in regulation is required iftrue e-trading is to get going. Mainly, regula-tors would have to be convinced that theirrecent reforms are in fact constraining a mar-ket that is a vital cog in the financial system.Markets in Financial Instruments Directive(MiFID) II’s pre-trade price-transparencyprovisions are just such a hindrance, in theeyes of many, and Basel III’s capital require-ments are another.


All in all, the changes described here seemlike a tall order for the corporate bond mar-ket. Most respondents to our survey expectonly a gradual evolution over the next sev-eral years. Barring a significant discontinuity,how will that slow change favor the variousplayers in the game? Primarily, we think itwill mean that independent operators ofquote-driven markets could find more suc-cess over the next five years. In particular, ofcourse, multi-dealer RFQ platform operatorsappear well positioned not only to capturethe lion’s share of the continued migrationfrom traditional voice- or messaging-basedtrading, but also to launch complementarycentralized crossing systems, as their rela-tionships with liquidity providers and thebuy side, and their installed base, couldprove valuable. But operators of multi-dealerRFQ platforms that are contemplating theintroduction of central limit-order booksmight want to think twice.

In spite of the pessimism about central limit-order book trading, exchanges could stillplay a role as potential operators of multi-dealer RFQ platforms, a not-so-distantcousin of quote-driven cash equities marketmodels, and crossing systems. Exchanges al-ready enjoy credibility from their operationsin other asset classes and from their reputa-tion for market neutrality. However, ex-changes would likely need to enter joint

ventures with key stakeholders such as majordealers in order to gain their support. On adifferent but related note, exchanges are bestpositioned to benefit from the exchangetraded fund (ETF) opportunity.

Bulge bracket dealers (and major buy-sideparticipants) that have recently invested intheir own trading platforms (for example,crossing systems) will likely need either to re-trench or accept subscale prospects. Thatbeing said, dealers and buy-side participantscan still shape the future of both multi-dealerRFQ platforms and crossing systems, as wellas other initiatives, by providing feedback,serving on user advisory boards and takingownership stakes (where appropriate) toalign their interests with those of operators.

Although the developments sketched abovedo not appear to favor operators of odd-lot/micro6 trading platforms, such platformswill remain part of the market structure, asthey provide liquidity to a segment (retail in-vestors) that cannot access larger-lot pools. Itis also conceivable that as institutional in-vestors sharpen their trading capabilities,they will increase their usage of smaller-lotplatforms to complement their core activi-ties. Furthermore, growth in corporate bondETFs may give them an additional boost, asauthorized participants may need to makesmaller trades to create and redeem shares.

Winners and Also-Rans

6 A micro trade is a trade with a valueless than $100,000.

E-trading may be slow in coming, but thisplodding evolution, along with post-crisisregulatory reform, is far-reaching, and al-ready fundamentally altering the business ofcorporate bond trading. The inventory-drivenmodel that market participants grew to relyon during the pre-crisis era continues to giveway to a new paradigm that mandates a setof skills and a technological infrastructureconsistent with a capital-light world. More-over, the steep sell-off of late may reset com-petitive dynamics in unforeseen ways.Consequently, all market participants wouldbe well advised to take steps now to ensureoptimal positioning for the range of direc-tions the market might take. The followingare actions that dealers, asset managers andoperators of multi-dealer RFQ platformsshould consider.

Dealers

In the capital-light era, dealers must thinkcarefully about how to make the most of thee-trading opportunity. Dealers should con-sider the following six moves—to the extentthey have not already done so:

• Align the organization. Dealers mustbreak down remaining resistance in thebusiness to electronification, and get peo-ple thinking more broadly about how to

use e-trading to compete better in flowproducts. This effort must start with un-ambiguous commitment from top manage-ment toward the e-trading business.Furthermore, dealers will need to adopt ane-trading governance model that balancescentralized governance of e-trading activi-ties and related infrastructure (such asconnectivity) with decentralized controlover product-specific elements (such aspricing, coverage and hedging strategies).Of course, dealers must take the costs andbenefits associated with each governancemodel—that is, the mix between central-ized and decentralized governance—intoaccount. Ultimately, strong leadership anda willingness to evolve in sync with themarkets will play a key role in determin-ing success.

• Centralize the trading infrastructure. Deal-ers should develop central pricing andtrade-processing infrastructure that servesboth traditional and electronic channels,thereby minimizing vendor-specific infra-structure, reducing time to market andlowering costs. Ultimately, this central in-frastructure should help dealers efficientlynavigate the market landscape, as well asmaximize crossing opportunities.


Preparing for the E-Trading Future

• Harness the data opportunity. Dealinghouses should consolidate their myriaddata interfaces (for example, interfaceswith internal systems, clients, market ven-ues and data providers) and store data ina centralized, accessible manner, ideally ina single “golden source.” This will helpfirms feed pricing engines more quicklyand effectively, and develop new tradingstrategies. Similar efforts to harness datacan drive other benefits. Dealers shoulduse their data to better inform sourcingand placement with clients and increasetheir share of “riskless principal” trading(in which they simultaneously find bothsides of the trade). Data sources that firmsshould cultivate include trade history,public disclosures of portfolio holdingsand RFQ history. Such a tracking systemshould be integrated with sales tools toimprove sales-force effectiveness. Superiordata management can also improve post-trade processes.

• Transform the sales force. Banks can pro-vide incentives to sales reps that rewardthem for shifting uneconomical trades andclients from voice to e-trading channels.Such incentives can be as simple as addi-tional sales credits for boosting e-usage by

uneconomical clients, and reduced salescredits for reps who continue to bookthese trades through voice. Motivating theright sales behavior will free reps to de-liver more valuable services such as ex-ploring client needs, generating tradingideas, providing market commentary or“color,” and offering counsel on complextrades. However, banks will need to at-tract, cultivate and retain a new breed ofsales talent with the skills to meet the de-mands of a new era—one far removedfrom the paradigm in which dealer inven-tory largely drove the sales function.

• Develop a content-rich e-interface forclients. The continued electronification ofcorporate bond trading will reduce oppor-tunities for dealers to engage in dialoguewith clients. As a result, dealers will needto work harder to differentiate themselves.While pricing consistency (that is, qualityof quotation) will remain paramount tosuccess in sales & trading, dealers that de-liver value-added services through chan-nels other than the telephone can captureadditional share. To this end, top-tier deal-ers should consider developing a clientportal—not to be confused with the sin-gle-dealer trading platform described ear-lier. Such a portal—facilitated by thecentral pricing and trading infrastructurereferenced above and linked to the param-eters upon which the client trades—couldintegrate pricing and other analytics,credit perspectives, insights into liquidityand market flows, and perhaps even tradeideas. In this context, the potential provi-sion of connectivity to all trading venueswould allow for the direct linkage of tradeideas and their execution in the market.Furthermore, the integration of this portal


Motivating the right sales behavior will free reps to deliver morevaluable services such as exploring

client needs, generating trading ideas, providing market

commentary or “color,” and offeringcounsel on complex trades.

with other client-facing technology couldprovide a front-to-back solution—frompre-trade to settlement. Last but not least,a client portal could help dealers serve thelong tail of smaller clients more efficiently.

• Reduce cost per trade. As noted above,firms can leverage e-trading to maximizestraight-through processing and cut post-trade costs. Doing so is critical in a marketwith smaller trades and increased tradingfrequency.

Asset managers

We recommend that asset managers revisittheir approach to corporate bond investingand trading by taking these actions:

• Reassess the investment decision-makingprocess. Asset managers should factor liq-uidity more prominently, and embed re-lated metrics more rigorously, intoinvestment decision-making in both pri-mary and secondary markets. Ultimately,these managers need to ensure they arepaid for the liquidity risks underlying thecorporate bonds in which they invest. Nat-urally, firms must equip their portfoliomanagement teams with the skills andtools necessary to operate in the post-crisiscorporate bond world—one in which theability to systematically price liquidity riskcould mean the difference between successand failure.

• Revamp the trading function to more effi-ciently navigate a fragmented market land-scape and economically transact in smallersizes. To begin, firms should decide whichmarket venues and dealers to engage withto execute their trading strategy. Theyshould also define their appetite and ap-proach for providing liquidity to the mar-ket as a price maker, and for disclosingtrading interest to a broader swath ofmarket participants.

With respect to technology, firms need toinstitutionalize front-office infrastructure.They should align the architecture offront-end systems (the order managementand execution management systems) withtrading specs to ensure that traders caneasily shift across liquidity venues, therebyminimizing missed opportunities. The buyside must also redefine middle- and back-office workflows and install systems thatprocess higher-frequency trading and cor-responding allocations to the underlyingfund vehicles in a scalable manner.

Organizationally, buy-siders with sufficientscale in corporate bond trading shouldconsider separating investment and tradingfunctions—if they have not yet done so—to assure sufficient specialization on thetrading side. Asset managers with multiplefund vehicles that invest in corporatebonds could consolidate corporate bondtrading desks across funds to maximizescale economies and crossing opportuni-ties. Furthermore, larger players might cre-ate an “odd lot” trading desk equipped totrade more frequently in smaller round-lot,odd-lot and micro sizes. Ultimately, firmswill need to assemble teams of traderssuited to trade in a world with a dwindlingnumber of inventory-driven dealers.


Ultimately, asset managers need to ensure they are paid for the liquidity risks underlying

the corporate bonds in which they invest.

Market operators

With multi-dealer RFQ platforms expected toremain the winning institutional corporatebond e-trading model for the next few years, itwill be interesting to see how current operatorsfend off new “e” initiatives launched bydealer/buy-side consortia, exchanges and infor-mation providers. While current multi-dealerRFQ platform operators are well positioned,they must continue to enhance their valueproposition to maintain that strong stance.

One way to do this is through the continueddevelopment of functionality that allows thebuy side to safely and anonymously syndicateits buy-and-sell interest with other buy-siders,as well as dealers. Platforms also need tomore effectively reach buy-siders and dealersthat have, or should have, a propensity totransact in those underlying issues.

Operators can also tap into nonbank sourcesof liquidity beyond the buy side; for example,by attracting independent proprietary tradingfirms. They can provide better tools and serv-ices to facilitate pre-trade activity and post-trade processing for market participants. Andthey can optimize connectivity through bettermanagement of their APIs.

More boldly, multi-dealer RFQ platform op-erators can innovate, potentially by develop-ing and launching a crossing system beforenew entrants can conquer the space. Toachieve success, it will be essential to gainearly buy-in from influential dealers and buy-siders. Operators can strengthen their rela-tionships with such market participants byinviting them to take advisory board rolesand, where appropriate, equity stakes.



Roger RudisuliPartner55 East 52nd StreetNew York, NY 10055+1 212 446 [email protected]

Doran SchifterSenior Expert55 East 52nd StreetNew York, NY 10055+1 212 446 [email protected]

Andy AwadManaging Director6 High Ridge ParkStamford, CT 06905-1327+1 203 629 [email protected]

Dan ConnellManaging Director6 High Ridge ParkStamford, CT 06905-1327+1 203 629 [email protected]

Contact

For more information, please contact:

corporate bond e-trading - mckinsey/media/mckinsey/dotcom/client... · 2020. 8. 5. · 2 corporate...

Documents