Why The US Treasury Market Needs Algorithmic Execution

By Alastair Hawker, Global Head of Sales, Quantitative Brokers

Algorithmic execution strategies are available for buy-side firms to automate execution in US Treasuries, improving performance and efficiency.

screen-shot-2018-10-01-at-4-31-10-pmAutomated trading, or specifically, algorithmic execution, of cash equities is prevalent to the point where execution without an algorithm is a rarity. In futures, use is not quite as extensive but it is well established. Even FX markets, which are greatly fragmented, have seen growth in execution algorithms in the last couple of years. And then there is fixed income, where it is almost non-existent in comparison.

Have a think about what the publisher of this magazine represents: the important FIX protocol that enables standardized messaging for order flow across the industry. It is used extensively for cash equities and futures, so why not cash Treasuries (USTs)? Why should most of the buy-side execution in one asset class be instant via a request-for-quote (RFQ)? Why should executing 1000 ten-year Treasury futures contracts be done differently to executing $100 million ten-year treasuries?

Buy-side firms have FIX connections to utilize many types of algorithms that enable them to access liquidity, reduce costs and enhance efficiency. However, although electronic trading of the US Treasury market is widespread between dealers and liquidity providers (the “inter-dealer” market), the buy-side is only at the beginning of the inevitable journey towards automation through execution algorithms.

It’s hard to imagine that USTs will not eventually be like cash equities and futures. We are not talking about the liquidity challenges of corporate credit markets here – the US Treasury market is an enormous market ripe for automation and the buy-side stands to benefit from an alternative to RFQs.

Fragmented market

Similar to equities markets and in contrast to futures markets, the electronic Treasury market is characterised by a fragmented landscape, consisting of numerous central limit order books (CLOBs) and direct pricing streams. The result is a complex structure that needs smart order routing (SOR) for successful navigation: aggressive orders need to be routed to where the cost is the lowest, achieving best price and minimum transaction fees. Passive orders are more challenging – they need to be divided between venues in such a way as to maximize the fill rate.

Recently, Quantitative Brokers (QB) completed a project on machine learning (ML) for SOR, which facilitates the intelligent aggregation of liquidity from multiple sources. It will be used to solve the problem of where to send child orders when there is a choice between different liquidity pools.

Current trading practices by the inter-dealer market participants have established a model for successful adoption of automation by buy-side firms. Almost 70% of all volume traded in the US Treasury market is conducted via electronic trading platforms, with over 90% of the nearly $200 billion traded daily (interdealer) executed electronically, according to research by Greenwich Associates.

Nevertheless, it is surprising how comparatively little buy-side execution in US Treasuries is automated and undertaken directly with electronic liquidity. RFQs still dominate – part of the reason is habit, but for various reasons there has also been a lack of investor access, or willingness to access, electronic liquidity.

Independent algorithms

Some of the challenges with access can be alleviated by firms that represent the buy-side as impartial agents, providing the connectivity and technology to aggregate liquidity and automate execution. QB is neutral and conducts no proprietary trading – its interests are entirely aligned with its clients.

QB’s suite of algorithmic execution strategies help buy-side clients minimize transaction costs, hide their footprint, and improve their productivity. These are available for on-the-run USTs, as well as futures markets.

QB’s flagship “Bolt” algorithm facilitates best execution across wide-ranging market conditions, benchmarked to arrival price. “Closer” provides optimal trading into the market close (settlement price benchmark); “Legger” intelligently manages legging risk for multi-leg orders; and “Strobe” attempts to capture the spread within a client’s defined time schedule, tracking a volume weighted average price (VWAP) or time-weighted average price (TWAP) benchmark.

Data

The key ingredient for successful automated trading is data. What matters for investors is understanding what means of execution is best for them. It is not just about requesting multiple quotes and trading at the best-quoted price. It is evaluating an entirely different way of executing and whether that is more optimal, taking into account pricing, anonymity and information leakage. Comprehensive transaction cost analysis (TCA) helps with this and is another part of QB’s client service.

There was also the introduction of TRACE reporting for US Treasury transactions in 2017. This has provided regulators with a more comprehensive picture of US Treasury market activity, but there is an unresolved debate about broader dissemination of this data to the public. Given the full transparency and resulting efficacy of listed markets, it is hard to understand how releasing this data would not be in the public interest. While this remains in discussion, the good news is that there is still enough data available for execution algorithms to work effectively. More data would always be welcomed, especially to help TCA benchmarking.

recommend to friends
  • gplus
  • pinterest