The Hybrid Multi-Asset Trading Desk

By Joseph Bacchi, Head of Multi-Asset Trading and Investment Operations, Acadian Asset Management

The multi-asset trader is the gatekeeper of execution quality and of the inherent economic and reputational risks that exist throughout the multi-asset trading process.

The mainstay of successful multi-asset trading is optionality.  This requires solid connectivity person-to-person contact with trust, understanding, mutual respect, and the use of direct market access (DMA) technology, with its benefits of cost efficiency and anonymity.  Success requires the understanding that the more tools and pathways are available, the better the trading outcome can potentially be.

As such, efficient trade execution requires a blend of automation and human agency: one cannot exist without the other.

The age of automation

Over the past 10-15 years, there has been a tendency to believe that the more automation is adopted, the better the process will inevitably become. This view had been – and continues to be – bolstered by a concurrent faith in the accessibility of Big Data.

It’s no doubt that automation has provided tangible benefits to the buy-side:

It has helped compress costs, explicitly by driving down commissions payable to sell-side brokers. It’s worth emphasising however that brokers were already reducing fees with their promotion of DMA facilities. Cost reduction was already embedded within the mindset of the industry, permeating the strategies and activities of most market participants. Automation (and Big Data) have increased the pool of liquidity and even uncovered new sources of liquidity to buy-side traders. The devising and adoption of sophisticated algorithms have streamlined some trading strategies, making them more robust and less prone to human error and inefficiencies. Automation has enabled greater anonymity for traders. Smart order routing and dispersing orders across different dark pools should, in theory, have plugged information leakage. But, in practice, algorithms and high frequency trading strategies have been able to identify trends and front-run orders – as have some unscrupulous human operators of those dark pools.

The human element

Experience, intuition, nimbleness, creativity: these traits have driven humankind forward from the cradle of civilization to the modern era.  They have spurred innovation from blacksmithing to industrialization to the electronic age.  These are the attributes that traders must possess and must hone every day—and they are traits that can never be replicated by a machine.

In the context of markets, a successful human trader can can understand the sentiment and dynamics at any given time or occasion. They have the ability to be flexible and to determine alternative ways to complete an order, such as using a derivatives or over-the-counter market, if they judge the conventional stock exchange to be inadequate.

Human traders are competitive, but they also cooperate to find a solution to ensuring the successful completion of an order. Camaraderie binds them together and unwritten rules governing behaviour and forging trust still matter.

Intuition is another important quality of a trader. Screen-based analysis is often insufficient compared with market knowledge and the understanding that allows a skilful individual or team to identify a trading profile and devise an execution strategy for a transaction. Cost, as indicated by a screen, is not always the best guide.

The limitations of automation

There are many problems that automation cannot solve, and there are circumstances where human trading skills are more effective.

An obvious example that institutional buy-side traders often encounter is accessing large blocks, especially of mid- and small-cap stocks. It can be a complex, nuanced process, requiring market knowledge built on experience, networks of personal contacts, and intuition.

Demand is relative to the supply of stock available in liquidity pools, and algorithms can’t discover and bring liquidity that simply isn’t there.

Smarter “new age” algorithms are being created to tackle this issue, notably in auctions at specific times during the day. They are useful tools for a trader, but they have shortcomings, so cannot be described as a “silver bullet”.

A trader exposes their order or position to other market participants, dealing spreads widen and the market becomes even more segmented, with brief seconds opened up for matching orders, which if they fail to match, means they are returned to a dark pool. Moreover, a focus on certain points where liquidity is expected to be available encourages passivity, with the risk that traders “sit” on their orders rather than try to work them.

Another objection to a reliance on automation lies in the factors or assumptions that are commonly imbedded in an algorithm. Take the widely used average daily volume (ADV) as an example. It might describe what has occurred for the past 30 days, but it cannot predict the volume available in the next 30 minutes. It is a guide at best, not an infallible basis on which to activate a trade.

Turning to other asset classes, automation in fixed income markets is well behind its adoption in equities markets.  While it is used in the most liquid government bond markets, for instance US Treasuries and the UK Gilts, it has made little headway in corporate fixed income, and least of all in the trading of tightly held high-yield bonds.

Partly this is because of illiquidity, but the lack of explicit cost is also a factor. There are no explicit dealing costs in bond markets, only the bid-offer spread, so there is less urgency about introducing automation.

Furthermore, most bonds, as well as many options contracts, trade over-the-counter, not on exchanges, where manual price discovery and order execution is only possible.

The hybrid model

The multi-asset trader has to be the gatekeeper of the process, understanding how to access different markets and knowing which are most suitable for particular orders. They must also compile counterparty profiles, be aware of exposures and be able to assess the quality of their back offices. Algorithms are unable to do these functions, because they lack the necessary nuance.

The reality is that for multi-asset trading, having access to both low-touch and high-touch trading strategies is necessary for execution quality.  Whether the matter at hand is new designs for liquidity sourcing in dark pools, or a structured product idea to help with customized exposures in less liquid instruments, multi-asset traders need both the efficiency that execution management systems technology offers, as well as the thoughtful strategizing that only a human can provide.

The decision as to when and how to use which approach should not be given away to the “inanimate” for the sake of ease or perceived risk mitigation.  The multi-asset trader must remain the guiding force, for they are the gatekeeper of execution quality and of the inherent economic and reputational risks that exist throughout the multi-asset trading process.  Knowing where you need to go is just as important as how you will get there.

The multi-asset trading desk has to be a hybrid to function effectively. It cannot be tied to a particular broker or restricted list of counterparties for all its orders, and nor can it be bound by a singular tool or methodology.