The Buy Side, Bank Algos and ECNs – It’s Complicated!

Most buy side organizations, particularly asset managers, have traditionally avoided prime brokerage relationships with their banks in the context of their FX trading activity. The reasons for this are understandable given the margin requirements and explicit cost of such an arrangement. Traditionally, the buy side has been limited to trading bilaterally with individual banks due to their ISDA relationships. The onset of algo platform offerings by large global banks opened the door, previously closed, to buy side institutions reaching ECNs and their multi-contributor liquidity.

By James L. Singleton
By James L. Singleton

The growth of this algorithmic execution alternative has been dramatic, particularly as regulators have enacted best execution mandates, and the buy side has been receptive to the benefits of executing their FX trades using more intelligent, machine-driven alternatives. But like all things new, the interaction of buy side execution style, specific algorithmic functionality (hint: there are many idiosyncrasies), and ECN liquidity characteristics creates a need for better understanding if these market developments are going to provide the benefits to all parties, which is of course the aim.

Algo products fall broadly into aggressive or passive strategies

Algo products fall broadly into aggressive or passive strategies

Best execution roll-out

The introduction and promotion of bank provided FX algorithmic products coincided with the build-up to MiFID II’s best execution roll-out. It also followed a period of enforcement actions that highlighted market behaviour that had been detrimental to customers’ FX trading objectives. To address these and other issues, banks began offering algorithmic products that offered access to third party liquidity options – ECNs, as well as the banks’ own liquidity. The algo products fall broadly into aggressive or passive strategies, and more often than not have been named after exotic animals. The algo providers charge a set price per million that covers all costs related to the execution. They also provide customers with transaction cost analysis of their trading activity. On a number of levels, these algo products deliver greater efficiency and choice for the buy side institution that is looking for best available price, transparent execution cost and post-trade TCA. But it is probably worth saying upfront that algo execution is not a simple default for meeting a customer’s best execution mandate.

What’s an Algo?

Simple question – simple answer: algos are trading programs that take orders and execute them in smaller pieces based on passive or aggressive strategies over some period of time. Without regard to their specific intelligence, algos will seek to find the best available price for the execution of each slice of the larger order. But as we scratch beneath the surface, there is more to think about. For example, an interesting characteristic of algos is the programs’ need to execute at the top of book price.

Algo platform offerings open the door to buy side institutions reaching ECNs

Algo platform offerings open the door to buy side institutions reaching ECNs

That consistent tendency might negate the benefit of tapping liquidity at pricing levels below the top of book – perhaps a missed opportunity at times. Given the intelligence built into algos and their various passive and aggressive strategies, how do algos interact with the specific characteristics of the ECNs connected to the algo platforms – and should you care?

Assessing Algo Platform Liquidity

Bank algo platforms connect to multiple ECNs that provide pricing which faces client orders. The bank offering the algo typically provides its own pricing as well. When the algo bank provider executes a client order using its own liquidity, that execution is “internalized.” All banks offering algos will disclose this internalization potential to their clients. MiFID II’s standard of best execution requires the market participant be informed and knowledgeable of all its trading alternatives.  So what does a typical buy side institution know about the various FX ECNs connected to its chosen bank algo platform? Our guess is little or nothing at all. What should it know? We would suggest a few, but very important details:

Is the ECN’s pricing firm or subject to last look?
Is the ECN’s liquidity streaming or subject to interest?
Is the ECN fully anonymous or not?
Does the ECN allow its liquidity providers to trade with each other?

We would hope that the buy side user of algos would know the basics about the ECNs connected to the platform and be curious how those ECNs interacted with the algo strategies being employed. If a buy side institution does not pursue answers to those questions, it is not reasonable to assume that it has followed a best execution protocol when using a bank algo.


As the title of this article suggests, the relationship among the parties engaged in the algo environment is complicated. And ironically, while the emergence of algo execution is an important leap forward in the best execution quest, elements of algo functionality and its market impact remain somewhat an open question. Suffice it to say that the sophistication of many bank algos is ahead of many buy side institutions’ knowledge base. The sales pitch for bank algos includes the benefit of reaching more liquidity faster. But that benefit can be undercut by the potential for signaling if the algo tries and fails to execute when facing last look liquidity. We are encouraged by the growth of algos and their benefits for the buy side. But the buy side will need to study how using different algos may impact their best execution process, particularly as they face ECNs with different rules and liquidity characteristics. It’s still early days but there’s plenty to think about!

Related Posts

Scroll to Top