What You Need to Know About Algorithmic FX Trading: Part III

In the last part of my series on algorithmic trading, I’ll be zooming in on the recent developments of this strategy in the forex arena. Before you read on though, make sure you’ve seen the first and second part of this article series!

As you’ve probably guessed, algorithmic forex trading has been on the rise in the past few years, as the appeal of automated trade signals and quick execution with minimal human supervision boosted interest and demand for “black box” systems. According to the “FX Electronic Trading 2014 – Global Trends and Competitive Analysis” report from Greenwich Associates, the usage of algorithmic trading has grown from 7% of forex market participants in 2012 to 11% this year.

In addition, Greenwich Associates predicts that the market share of algorithmic trading will reach 18% of forex traders by the end of 2014. And that’s just on the retail level!

More money in "black boxes"?

More money in “black boxes”?

The adoption of algo FX trading systems among hedge funds and institutional trading firms is projected to be significantly higher, as these market participants enjoy larger capital and can be able to make the most out of higher trading volumes. “Hedge funds using algos execute even more volume through them, over 50%, and we expect that to only increase,” said Greenwich Associates Market Structure & Technology head Kevin McPartland. Other studies indicated that high-frequency trading systems account for roughly 75% of volume among larger funds.

Of course the development of algo systems throughout the years and across various financial markets has been riddled with a few problems here and there. A couple of years back, the opening of the BATS (Better Alternative Trading System) Global Markets exchange – an ECN stock exchange with trading systems based on trade bot, a computer program performing algorithmic trading – set off a market glitch among several equities.

“Black box” systems have also been blamed for a flash crash that occurred in May 2010 when Waddell & Reed Financial, Inc. implemented an algorithm that triggered a feedback loop among its high-frequency trades. No need to go into the nitty-gritty of this blunder, but you should know that it sparked the largest intraday dip in the Dow to the tune of 998.50 points!

Back in 1998, a highly-leveraged quant hedge fund trading in the fixed-income market got instantly blown out when Russia announced a default on its bonds. The government had to scramble for a bailout plan for this firm in order to prevent contagion and a full-scale debt crisis from happening.

In the forex market, several firms have already been placed under scrutiny for being suspected of using algorithmic systems that are designed to manipulate WM/Reuters exchange rates. Industry watchdogs are also looking into the possibility that traders may have colluded with dealers in order to push forex rates in their favor during the 60-second windows before the benchmark rate are set.

Suffice to say, this part of the financial industry could be in for even more promising developments in the coming years but this merits closer regulation as well. I’ll be sure to share any updates with y’all so stay tuned!

  • Very interesting article Espipionage. The landscape is indeed varied. There are different classes of algorithms, including HFT, program trading, etc.
    Interesting how fast the retail portion of automated trading is projected to increase in forex. Latest data and source for retail very interesting, thank you.
    On the institutional side of things, I think the volume traded by computerized algos was already 45% in 2010. After 4 years we definitely have to expect even more than 50%.
    Technology and regulatory changes (e.g. RegNMS and MiFID) have created market fragmentation and conditions where algos can thrive for, at least, the three following reasons:
    1) market is now a network of computers therefore is simpler and more natural for orders to be generated through algos running on computers
    2) brokers have an incentive to use “smart routing”, based on computerized algos, to satisfy the obligation to obtain the ‘best execution’ and the rule of ‘order protection’ (rules, unfortunately, too often bypassed by HFT in dark pools and other venues)
    3) multiple market platforms and venues generates opportunities for arbitrage for those who can move quickly between the trading venues. Arbitrage can only be implemented with very fast operations, again requiring algos.

    Some of the changes that we may see in the future are algos embedded in hardware, rather than software, and trading logic that is designed with the help of computers!
    It is going to be very interesting.

    Happy trading
    Giuseppe Basile, FibStalker

  • Pi-Pi-Piggy

    Hi Everyone

    I am a newbie and currently busy going through the grades of the school of pipsology. I would like to know why I shouldn’t just jump on the back of a successful program and take the ride, instead of going through the months of training and creating a method plan.

    • Cyclopip

      Hey Pi-Pi-Piggy! Great name and good question! Putting your money into an automated trading program is no guarantee of success for many reasons: the code may not be updated or targeted for the current environment, the advertised return may not be truthful, the risk management parameters may not be inline with your comfort level, and not being the original programmer of the algo means you cannot adapt the program to the shifting environment. Just some things to think about before investing your hard earned money into something you know nothing about.