The Rise of the FX Algo Wheel

Algorithmic trading has proliferated across global FX markets over the past decade. Today, roughly 20% of the institutional foreign exchange trading volume is now executed through algos. FX algo usage is following that of the equities market - where algos currently account for more than half of all equity trading volume. So what’s behind all this algo growth?

The Rise of the FX Algo Wheel
By Bob Afett

FX Algo usage is being driven by two primary factors: lower execution-cost and higher execution-quality. Both have always been used to improve relative returns. But this push has taken on new urgency, following a raft of recent Regulations around the globe. 

The past few years have witnessed legislation such as Dodd-Frank, EMiR, MIFID2 and Basel III. All sprouted from a desire to reduce systemic risk. Whether or not they’ve achieved their aim is unclear. But they have accomplished one thing; they’ve now made the Buyside institutional trader responsible for their own execution quality. 

Formerly, the Buyside trader would pass their FX orders to the dealers and hold them accountable for “best ex.” Now, the Buyside trader is on the hook for WHERE, WHEN and with WHOM their FX was traded. Under these new regulations, best-ex is no longer the dealers’ problem. The Buyside must now illustrate that they’ve selected the best broker and algorithm - and be able to prove it.

This new Buyside accountability extends beyond the FX markets alone. But what was once just a technology story has now become a Regulatory one. Along with the increased focus on transaction cost analysis and you begin to see why algo us is growing in the FX markets.

But here’s the tricky part.

The Buyside must now undertake “regular and effective monitoring” of their trading tools. This effectively imposes governance and testing obligations upon any investment firm using any algorithms. Thus, the managers themselves are legally liable. This means they must now divert time and resources towards each algorithm they use. 

Unless they can solve this headache with technology….

Big Wheel

Enter the Algo Wheel.

Software is now efficiently solving these Buyside problems of execution-cost, trade-performance and algo-monitoring.   How?  Through data, of course!

The easiest way to understand the FX Algo Wheel is to view it as more of a process than a product. It utilizes software to automate routing decisions, based upon a series of data analysis. The Wheel tracks and measures execution at a granular level. Guided by data, rather than a human, the order-routing is then directed by relative performance. The Wheel’s data can further assist providers in improving their own FX performance going forward. Think of it as a collaborative carrot and not a punishing stick. 

Performance driven trading

The industry term for this new approach is “performance driven trading.” But this isn’t just some snazzy marketing pitch. It combines elements of the quant world with traditional “old school” trading. Once the human has selected their FX strategy (% of Volume, TWAP, Passive, etc.) the Wheel then directs the order to the  broker with the statistically best execution. The best analytical performance drives the trading decisions. 

Simple! ...Or is it?

Like most things in life, the devil is in the details. In the case of Algo Wheels, defining the “statistically best“ can become complex. And even more so with FX trading. Here is why.

To determine the “best” in anything, you must first establish a baseline of expected performance. You can’t know if the trading was better than normal until you first define “normal.” 

Challenges

Comparing relative performance across any data universe is challenging for several reasons. First, it requires accumulation of a “statistically significant“ data-set to compare against. Second, that data must be “normalized” across all providers to conduct an apples-to-apples comparison. Finally, accurate analysis must also account for order constraints such as broker-restrictions and participation-limits.

With data analysis, more is always better. Unfortunately, no single firm ever produces a large enough sample-size for truly definitive conclusions. Conclusions are more akin to the “best right now“ rather than an unequivocal “best.” 

Obviously, currency trading happens in pairs. Unlike the stock market, ForEx requires the buying and selling of multiple currencies simultaneously. So another challenge arises from the liquidity dynamics of the market itself. The vast majority of the volume in currencies happens in just 18 pairs. As you know, FX trading is most heavily concentrated in the big eight:

  • US dollars (USD)
  • Canadian dollars (CAD)
  • New Zealand dollars (NZD)
  • Australian dollars (AUD)
  • Euros (EUR)
  • British Pounds (GBP)
  • Swiss Francs (CHF)
  • Japanese Yen (JPY)

But in the face of these liquidity challenges, an FX Algo Wheel can make the universe of less liquid currency pairs, much easier to manage. And by quantifying trade performance across brokers, it enables steady improvement over time. Perhaps most importantly, it also allows the Buyside trader to clearly explain and justify their “best execution practice“ to both clients and regulators. How?  The same way you remember from school!

Report Card

FX Algo Wheels generate a quarterly report (or monthly) of each broker’s trading performance. 

The analysis typically compares against a “peer database” comprised of six months of the most recent data. That information should account for market factors such as relative volatility and volume. “Report cards” should also explain the cost-estimates used in their analytics modeling. The most advanced reports can enable a comparison against specific subsets of data, for more customized insights. 

Remember, the goal is to create discussion points to help the FX brokers improve their execution. Instead of simply hearing their performance is “terrible,” those brokers will now have a more concrete starting point to begin tweaking their approach. 

FX Algo Wheels generate reports of each broker’s trading performance.
FX Algo Wheels generate reports of each broker’s trading performance.

Conclusion

The concept of a broker-neutral randomization tool is not new. 

But prior to the Algo Wheel, such “switchers“ and “allocators“ lacked the “big data“ analytics. Recent advancements in both data science and routing technology are changing the business of trade execution. This discipline will ultimately lead to machine learning, through we are not there yet. Currently, no full-scale, A.I. Algo Wheel exists in any marketplace. 

The competitive field of FX Algo Wheel providers is still wide open as none of the emerging offerings have yet established any clear dominance. Whether the locus becomes the OMS vendors, the brokers or some other technology provider is still not established. But the evolution is accelerating as the Wheel proves its merit, outside of currencies.

Exciting as this may be, the promise of automation is always a double edged sword. 

All algos face the unspoken concern that automated trading products will lead to long-term erosion of the human trader. But FX Algo Wheels are no replacement for human expertise and judgment. They are merely an efficiency tool for the “no-brainer” workflow. The truth is that Algo Wheels are not a “silver bullet” and blind faith in any automation may get you killed. So the best use of the FX Algo Wheel still lies in the handling of the plain vanilla trading. This will allow you to better focus on the complications of fudge-ripple.

And who doesn’t love a good ice-cream sundae at their desk while the machine handles the work?...