MarketAxess Holdings, Inc.
Jan 18, 2017
PDF
Add to Briefcase

Big data touted as next key to unlock fixed income trading

Reuters – Tom Porter
January 18, 2017

Data analytics firms could be one of the prime beneficiaries of poor liquidity in fixed-income markets, as investors look for new tools to help them trade bonds in meaningful size.

Electronic trading of corporate bonds is growing - fueled by the rise of platforms matching buyers and sellers without banks acting as intermediaries - but over 80% of the roughly US$6trn traded annually in the US is still matched and executed over the phone or via instant messenger, according to research by Greenwich Associates.

"While electronic trading growth in corporate bonds tends to make the headlines, improving the human-aided liquidity-seeking process is an increasingly huge part of the story," said Kevin McPartland, head of research for market structure and technology at the firm, which polled around 2,000 US and European institutional fixed-income investors.

Big trades, big data

Increasing numbers of institutional investors are convinced of the benefits of e-trading platforms, according to Greenwich. Over half of those polled said they were planning to use a venue on which they were not currently active in the coming year.

Over 80% of credit investors in the US and Europe still feel that reduced market liquidity is impacting their ability to implement their investment strategy.

While e-trading volumes are increasing, the new entrants have so far failed to solve the biggest problem facing corporate bond investors, which is their ability to transact in sizes over US$2m.

In February 2016 the average size of inquiry from the buyside to the sellside seen by trading software provider Algomi averaged roughly US$8m, suggesting investors are seeking help in executing larger trades.

Algomi is one of the better-known players in an industry that Greenwich Associates calls "liquidity intelligence", which aims to provide buyers and sellers with better data on the exact location of liquidity scattered around the bond markets.

"While capital requirement rules continue to squeeze the ability of dealers to provide price information on illiquid bonds, technology is playing an important role, allowing banks to securely and intelligently harness data to make valuable financial trading connections," Algomi chief executive Stu Taylor told IFR.

New firms offering 'big data' solutions to bond traders continue to emerge.

London-based Mosaic Smart Data was set up by former Deutsche Bank rates trader Matt Hodgson. Its software offers "predictive data", tracking bids and offers in the market so it can point traders to where there is liquidity available in certain bonds.

Mosaic was in advanced talks with a tier one US investment bank to install the software towards the end of last year.

'Virtual' balance sheets

While market participants broadly expect the majority of fixed-income trading tech firms to fall away or be acquired by larger rivals, it is looking increasingly likely that bond trading in the future will require the use of multiple venues and tools.

Over 90% of an institutional investor's electronic trading is still done via their primary platform, according to Greenwich, but the average number of venues used has risen steadily from 1.2 in 2014 to 1.5 in 2016.

The business is dominated by Bloomberg and MarketAxess, which have 89% market share between them, according to Greenwich data, but firms such as TradeWeb, which is majority-owned by IFR's parent Thomson Reuters, Liquidnet, TruMid and Electronifie are expected to gain traction in the coming months.

Opening the pool of potential buyers and sellers in the form of the all-to-all trading model, championed by the likes of MarketAxess and now Liquidnet, is just one part of the solution to illiquidity in fixed income, said McPartland.

Mining data to create "a view into the location of every bond in the world" is arguably a more impactful approach over the long run, he said.

"If Uber can upend the taxi business without owning any cars, it might be possible for a complete data set with the right intelligence to create a virtual balance sheet without having one of its own."

Read more: http://www.reuters.com/article/idUSL5N1F81Z9