Darkness Rising: Can Dark Pools Cure What Ails Credit Trading?

As 2014 drew to a close, a multi-strategy hedge fund in the US Midwest was puzzling over the performance of its credit strategies. The firm had been deeply pessimistic about credit markets for some time, and told its clients as much in a series of gloomy investor letters. Despite this, its credit strategies were consistently beating expectations. Now, some of its investors wanted to know why.

"We were scratching our heads," says a partner at the firm. "We eventually realized we had slipped into the role of the prop desks. We had become liquidity providers in the credit markets. Instead of paying the spread, we were collecting it. Trading had become a source of income. We had to explain it to our investors."

The future of credit markets may depend on more firms following this inadvertent trailblazer. Bank market-makers are being sidelined by tough new capital rules, and there is widespread agreement ─ among dealers, investors and regulators ─ that the only way to get liquidity flowing again is to rope in the buy side.

But how? The credit market is a vast galaxy of different securities, most of them dormant for weeks or months at a time; transactions of less than $1 million represent 85 percent of the flow, but only 15 percent of the volume, and with trading sporadic at the best of times, participants are paranoid over keeping their intentions secret.

A Bigger Elephant, a Smaller Keyhole

“If I engage a sell-side trader, nine times out of 10 they can’t do it in the size I want. So, Often you have a good idea and no ability to execute. With a dark pool, I can try and access liquidity, and if I don’t get anything, no one will know what I was trying to do.” Former bond trader, New York hedge fund

Trading credit in size has never been easy ─ one trader compares it to squeezing an elephant through a keyhole ─ and it has become more difficult in recent years.

The credit market ─ the elephant, in this example ─ is larger than ever, with the total value of US corporate bonds outstanding having grown 50 percent from $5.2 trillion in 2007 to $7.8 trillion at the end of last year, according to data from the Securities Industry and Financial Markets Association. The keyhole, meanwhile, is getting smaller, with dealer inventories shrinking from $250 billion at the end of 2007 to $57 billion, according to data from the Federal Reserve Bank of New York-a drop of nearly 80 percent.

As a result, dealer inventories of corporate bonds account for less than 0.75 percent of the market compared with around 5 percent in 2007-with more than 99 percent of debt outstanding controlled by asset managers and end-investors.

The decline in dealer risk capacity calls for a new market structure, some argue, one where the buy side plays a larger role in liquidity provision. Large asset managers such as BlackRock and Pimco seem to agree, and a host of new venues have sprung up to capitalize on the interest in all-to-all trading.

"The dealers were the stabilizing force in the credit market," says a credit trader at a New York-based hedge fund. "They were willing to commit capital to facilitate risk transfers. They don't do that anymore, so the market has to evolve."

Mix of Platform Types

The solution will probably require a mix of platform types and trading styles, participants say. At least 15 platforms are now up and running, or preparing to launch, offering everything from exchange-style limit order books to dark pools. If the latter take off, it would be a first for credit markets ─ and perhaps not what regulators have been hoping to see ─ but the seven buy-side firms interviewed for this article were receptive to the idea.

"Dark pools definitely make sense as a concept in the credit world," says Michael Nappi, senior investment-grade trader at Eaton Vance in Boston. "Previous efforts at creating them for bonds have failed because they were single-dealer supported. The market really doesn't like that model. But if you have an independent party doing it, and they do it the right way, it could have legs."

A former trader at a large, New York-based hedge fund also sees potential: "There have been lots of cases where I might have an idea about a certain company and decide I want to own the bonds in a particular size. If I engage a sell-side trader, nine times out of 10 they can't do it in the size I want. So, often you have a good idea and no ability to execute. With a dark pool, I can try and access liquidity, and if I don't get anything done, no one will know what I was trying to do."

As things stand, liquidity provision in credit is handled almost exclusively by the 15 or so major dealers ─ with investors either calling salespeople at multiple banks to ask for prices or completing a request-for-quote (RFQ) on one of the established electronic trading venues ─ MarketAxess in the US and Bloomberg or Tradeweb in Europe. Around 85 percent of credit trading in the US is done on the phone, with electronic RFQs accounting for most of the remainder.

Changing the Structure

The first serious attempt to change the market structure came in 2012 when MarketAxess launched its all-to-all Open Trading initiative.

That effort won strong backing from BlackRock and resulted in the introduction of two new trading protocols: Market Lists, which allows users to send anonymous RFQs to all members of the platform; and Open Markets, a bulletin board-style order book where dealers and clients can post their orders and axes for all to see.

After a slow start, which saw only 13,000 all-to-all trades completed in 2013, Open Trading is gathering momentum, with more than 77,000 trades last year ─ accounting for around 7 percent of all credit transactions on MarketAxess. It is also attracting new liquidity providers to the market. More than 300 firms responded to Open Trading RFQs in the fourth quarter of 2014, up from around 150 during the same period in 2013.

That is a big step-up from the roughly 90 dealers that act as liquidity providers on a name-disclosed basis on MarketAxess' traditional dealer-to-client RFQ platform. Even assuming every dealer was responding to all-to-all RFQs, it still means more than 200 non-dealer firms ─ traditional asset managers, hedge funds, exchange-traded fund (ETF) market-makers and proprietary trading shops ─ provided liquidity to their peers on an anonymous basis in those three months.

"Open Trading allows these firms to act as price-makers and liquidity providers, and to do it opportunistically in a way they never could in the past," says Richard Schiffman, Open Trading product manager at MarketAxess in New York. "These new liquidity providers are replacing some of the $200 billion of lost dealer liquidity. They're all small compared to the dealers, but collectively they're providing a meaningful amount of liquidity."

Still, not everyone is sold. The average trade size on the platform is just under $500,000 ─ and critics point to this as evidence that the all-to-all RFQ model is ill-suited for trading so-called "round lots" of more than $1 million and block sizes exceeding $5 million.

Schiffman does not dispute this. "We do see round lot and block size business in the types of trades that are not hyper-sensitive to information leakage. But some asset managers are anxious about using competitive RFQs and letting their intentions be known on longer maturities in sensitive credits. That potentially requires some new protocols, and we're working on that along with lots of other people," he says.

Clob Alternative

The traditional alternative to RFQs is an exchange-style central limit order book (Clob), where market participants stream anonymous orders into a single pool and compete to offer the best prices. Clob trading already exists in the credit markets.

UK-based Vega-Chi launched the first Clob for US and European high-yield bonds in 2012. But the platform has struggled to attract liquidity from dealers, which see little reason to support anonymous trading venues. Vega-Chi was acquired in March by Liquidnet, which operates a dark pool for equities. NYSE Bonds and MTS Bonds, a subsidiary of the London Stock Exchange Group, also offer Clob trading for credit-and have also failed to generate significant volumes to date.

Some dealers offer streaming prices and Clob-style all-to-all trading on their single-dealer platforms. The UBS Price Improvement Network (Pin), for instance, is notable for allowing clients to make prices on the platform, which are displayed alongside the bank's own streaming two-sided prices. "That's where it really gets interesting," says Kevin Arnold, head of foreign exchange, rates and credit in the Americas for UBS in New York. "We stand there as riskless principal ─ the firm's credit rating and balance sheet is behind every transaction ─ but the price is derived from a whole host of liquidity providers."

Firms such as UBS are trying to capitalize on the benefits of all-to-all trading, while maintaining their client relationships. Arnold says the bank's strategy reflects the realities of market-making in a post-Basel III world. "The rules have changed. Dealers are not the sole arbiters of liquidity any more. They can't be. They're not allowed to be," he says. "The buy side has become the lifeblood of the system. If some dealers still think they're the only ones that can put a price on something, they're fooling themselves, and the market structure will change regardless."

HSBC has come to the same conclusion. Last year, it launched HSBC Credit Place, a multi-participant platform where institutional clients can place resting limit orders anonymously. Much like UBS Pin, client prices are displayed alongside HSBC's prices.

"It's a platform where HSBC shows prices from its trading desks within credit globally alongside the liquidity available from clients. The idea is that if a client in Europe is trying to sell a bond, someone in the US or Asia may want to take the other side because they're managing to different metrics," says Soteris Manderis, director of electronic trading at HSBC in London. "Everything is done anonymously, with HSBC in the middle as the counterparty to all transactions."

Carving a Niche

Single-dealer credit platforms may yet have a place in the new market structure. UBS Pin, for example, has carved a niche for itself in single-name credit default swap trading, according to market sources. HSBC signed up 78 institutional clients and received around $20 billion of orders on Credit Place last year, according to Manderis.

However, critics argue that single-dealer platforms cannot solve a market-wide problem and may fragment liquidity, hampering efforts to consolidate trading capacity at impartial, multi-dealer venues.

That has spurred some dealers to throw their weight behind an effort by Tradeweb to break into US credit trading, and challenge the dominance of MarketAxess. The Tradeweb US Corporate Bond Marketplace was launched last October and offers live, streaming and executable dealer prices for round-lot investment-grade trades of more than $1 million, alongside a dealer-to-client RFQ mechanism.

Tradeweb's streaming multi-dealer prices and click-to-trade functionality are a leap for the credit markets. The platform features 10 of the largest dealers streaming two-sided prices for around 300 of the most liquid high-grade bonds. Tradeweb insists the prices are firm, and clients are told their orders will be filled at the displayed prices with 95 percent certainty.

Despite the novelty of firm, streaming multi-dealer prices, Tradeweb's foray into corporate bonds met with a tepid response among asset managers. "The fact we have a streaming market for the most liquid investment-grade bonds is great ─ but those aren't the trades that we need help with," says Nappi at Eaton Vance.

Billy Hult, president of Tradeweb, says that he fully understands the frustration. "If we could snap our fingers and solve for the really illiquid bonds, where the heat is, we would do that. But you have to start in a place that will get the community comfortable, and grow and expand from there. We're making significant progress toward the electronification of the corporate bond market, and we're going to listen to the buy side and evolve the platform in a way that creates real value," he says.

Tradeweb may also have some work to do on the sell side. Rank-and-file credit traders in the dealer community are hardly ecstatic about streaming prices to the platform.

"The US has always been a runs-and-quotes-based market ─ streaming prices is an alien concept," explains a New York-based credit trader at a large bank. "We were supportive, but the volumes and client interest haven't been huge. It's a lot of work to keep those prices live and updated at all times for not a lot of reward."

Tradeweb's struggles do not augur well for the truly anonymous exchange-style Clobs, such as Vega-Chi, NYSE Bonds and MTS. "The difficulty with streaming anonymous prices to a Clob is that you're theoretically providing liquidity to your competitors. That's always been a concern among dealers. Doing attributed trades with clients can generate added value to the franchise. But streaming prices to a Clob, and getting hit or lifted by your competitors, is not an efficient use of resources," says the New York-based credit trader.

Ultimately, this all comes back to what Amar Kuchinad, chief executive of bond trading platform Electronifie calls the "85/15" problem. "If you look at the data, 85 percent of corporate bond trades are less than $1 million in size, but they account for less than 15 percent of volume," he says.

Much of this 15 percent is already being traded on MarketAxess' electronic RFQ platform. Shifting a portion of it to an all-to-all model will help at the margins by freeing up dealers to focus on the more difficult trades, but it still leaves the market relying on banks to facilitate 85 percent of corporate bond trading by volume.

"The problem that we're trying to solve is the lack of immediate liquidity for larger transactions. Investors are finding it hard to source liquidity when a trade becomes too large for any individual dealer's balance sheet, and the efficiency of that execution is hampered by the leakage of information, so the market impact can be relatively significant," says Kuchinad.

This feature first appeared in the April 2015 issue of Risk magazine. Next month Waters will publish the concluding part of this article.