Insights

Bond Algos Tap into ETF Liquidity and Efficiency Gains

March 11, 2019 | By: Ivy Schmerken

By Ivy Schmerken, Editorial Director

Algorithmic trading has picked up steam in the corporate bond market as fixed income trading desks appear to follow the playbook of equities. But skeptics contend this is unlikely, arguing that bonds are quite different and will undergo their own electronic transformation.

One of the world’s largest fund managers has developed a pre-programmed system that executes buy-and sell orders with minimal handling by its traders, according to a speaker at Tabb Group’s “Survival of the Fittest” fixed income conference in February.

With a goal of “zero to low-touch trading” the asset manager follows pre-set rules and instructions to execute each order.

Corporate bond algos allow the buy side to respond to thousands of prices –request for quotes (RFQs)– sent by the brokers and then automatically execute on certain parameters.

“Right now, the way it works is to have a set of bots that are set up to deploy and scan large RFQs from the dealers in the market and automatically respond to RFQs,” said the buy-side speaker on “Bond Algos – Equities Redux or Can Liquidity be Manufactured.”

Efficiency gains and the search for opportunistic liquidity sources are driving adoption on buy-side fixed-income desks.

Banks have pulled back from responding manually to quote requests.

Some bank desks reportedly get thousands of requests for quotes per day. Brokers can cut costs by pricing bonds through an algo and the buy side can trade debt more efficiently.

However, on the credit side, liquidity is still challenging and the trades that are executed automatically are retail size or odd lots.

Whether there’s enough liquidity on the credit side for these algos to work is the question, said Larry Tabb, founder and chairman of Tabb Group, who moderated the panel.

Only 20% of corporate bonds are traded electronically partly owing to the complexity and numbers of bonds. Unlike in equities where each company has one common stock, each issuer may have many bonds with different maturities, yields and characteristics. Despite these obstacles, significant progress has been made for odd lots – trades less than $1 million.

“We are in the early stages of algorithmic trading,” said the buy-side trading technology executive whose firm is emphasizing efficiency, technology and innovation

“We set up an algo to automatically respond to RFQs with certain parameters, such as which part of the curve do you like, which bonds do you prefer, which bonds are attractive, which names are very active by our investment management teams,” said the buy-side executive.

“This basically allows those algos to run and source opportunistic liquidity for our investment portfolios,” said the buy-side speaker.

Internally, the asset manager has developed a user interface that allows everyone on the investment team to access the list of parameters and modify them so it’s transparent.

A credit research analyst or a portfolio manager or a trader can go into the system, update their market views, and update which credit structure they prefer in bonds on their list. The list gets refreshed every day or throughout the trading session.

Sell Side Automates Liquidity Provisioning

Banks have developed their own algorithms for pricing odd lots, which is considered a “fertile ground” for expanding electronic trading in fixed income, reported the Financial Times in “Bond trading algos find fresh hunting ground.”

Credit Suisse, Goldman Sachs and Morgan Stanley are among the banks that have reportedly developed corporate bond algos, reported the Financial Times in June of 2018. Citing data from electronic fixed income trading platform MarketAxess, the number of banks using algos to price odd-lots doubled from four to eight, wrote the FT.

On the panel, a sell-side market-structure expert said the bank’s corporate bond algo is pricing 10,000 CUSIPs in investment-grade and high-yield bonds. It takes in different inputs, formulates a price and then performs automated risk management. The goal is efficiency for its traders in pricing bonds. It “takes the burden away from round-lot voice traders [having] to respond to all RFQs that come through electronically,” said the sell-side market structure executive.

New Players in Algo Space

Meanwhile, demand for low-touch, automated investment-grade credit-trading has drawn new entrants such as non-bank liquidity providers.

One proprietary trading firm’s CEO speaking on the panel has rearchitected traditional trading workflows with a systematic and quantitative approach to electronic market making and risk management.

Launched two years ago, the proprietary trading firm is providing liquidity through corporate bond algos to multiple venues and protocols. It’s seen an uptake from clients appreciating the immediacy of click-to-trade.

In addition to auto-quoting and responding to streaming prices, it set out to automate “the whole lifecycle of data modeling, processing, data acquisition, portfolio allocation, and risk management,“ said the firm’s CEO, who estimates that 80% of the workflow is automated.

While the prop trading firm still has a human trader monitoring the trading book, the firm’s process is on auto-pilot. “It’s running by itself, it’s processing trades, it’s risk managing, it’s getting in-and -out of positions automatically.”

ETFs Provide Liquidity

Experts maintain that electronic trading and corporate bond liquidity is closely tied with exchange-traded funds or ETFs. “There is a bifurcation of liquidity between bonds that are components of ETFs, which are based on an index of bonds as opposed to non-index bonds,” said the proprietary trading firm CEO. “The bid-offer is compressed more substantially” for the bonds that are part of an index tracked by ETFs,” he said.

Popular ETFs like HYG and LQD issued by Blackrock are contributing to corporate bond liquidity, panelists suggested.

“Authorized participants involved with create/redeem baskets in the underlying bonds contribute to liquidity every day,” explained the CEO.

“In that sense, ETFs are creating a more efficient market,” observed Tabb, pointing to the standardization of corporate bonds in the index and the creation/redemption process for ETFs.

Bond Algos & Portfolio Trading

Buy-side firms are also utilizing algos to submit portfolio trades to the banks. With automated pricing via credit algos, there has been a speed up of liquidity around the index products leading to portfolio trading of corporate bonds. “Portfolio trading is not a new phenomenon. What makes it different now is that we have an algorithmic, automated pricing tool to price really, big portfolios,” said the sell-side market structure executive. “When clients have a large outflow, now we have an algo that helps the trader price it and book it within an hour.”

But this wouldn’t happen if not for broker dealers allocating capital to the algo book. “It’s not just efficiency, but specific capital is allocated to our algo book,” said the sell-side executive.

Looking Ahead

While institutions are gaining a comfort level with bond algos, they are evaluating bond prices and whittling down the list of bonds to those with liquidity.

Algos will continue to evolve. However, for a large accumulated position, the buy side still needs to pick up the phone, said the sell-side market structure panelist.

“It will never replace the client relationships and the voice trading,” said the investment bank executive. “To us we see it as an additional kind of trading protocol that we offer to our clients and their tool kit for execution.”

Experts refute the notion that algo trading in corporate bonds has advanced to the point of electronic trading in equities.

“The sell-side’s algos are more like auto-quoting bots that stream prices or respond to RFQs,” said Tabb in a follow-up interview. But “they are not like algos that break up a larger trade into smaller pieces and VWAP them,” he said. Also, he doesn’t think there are many market making bots that enter limit orders into live limit order books. For these reasons, the corporate bond algos are not on par with algos in equities, at least not yet, he said.

But as the machines begin to replace the human traders from pricing the small lot orders, could the banks have bigger goals in mind?

“Bond traders see scope for slicing the big corporate bond orders into smaller chunks creating the opportunity to sell them algorithmically in bursts rather than calling up banks for quotes or messaging banks for quotes,” reported the FT article.

It remains to be seen how these developments will unfold. But one thing is clear, the buy side is focused on efficiency and that has motivated the development of bond algos. Time will tell how far they go.