CEO CHAT: David Mechner, Pragma Securities
And the algo trades on…
Algorithmic trading has been around since the early 1970s, starting with the introduction of the New York Stock Exchange’s “designated order turnaround” system (DOT, and later SuperDOT), which routed orders electronically to the proper trading post. From there program trading evolved – again being entered with the aid of a computer. From there the equity markets with fully electronic execution and similar electronic communication networks developed in the late 1980s and 1990s. This increased market liquidity led to institutional traders splitting up orders according to computer algorithms so they could execute orders at a better average price.
Then at the turn of the century in 2001 a team of IBM researchers published a paper at the International Joint Conference on Artificial Intelligence where they showed that in experimental laboratory versions of the electronic auctions used in the financial markets, two algorithmic strategies (IBM’s own MGD, and Hewlett-Packard’s ZIP) were born. From there, algorithmic trading evolved further with new strategies such as VWAP, TWAP and others with technology vendors creating myriad iterations.
One of the early algorithmic trading pioneers is David Mechner, Chief Executive Officer at Pragma Securities. Pragma has been at the forefront of algo technology and recently John D’Antona Jr, editor of Traders Magazine, caught up with Mechner and talked about the current technology environment, adoption in new asset classes and innovations and how regulation can affect electronic order flow.
Traders Magazine: How would you describe the state of algorithmic trading in the equities and FX markets?
David Mechner: Algorithmic trading is deeply ingrained in the equities markets, accounting for approximately 80% of volume. In addition, over the past year MiFID II’s unbundling of research from trading, and its best execution requirements, has led to increasing intensity of focus on execution quality and transparency. This focus is driving buy-side traders to spend more time understanding how brokers’ offerings work and perform, which in turn is driving brokers – at least those who want to maintain a robust trading relationship with their customers – to invest more in developing a high-quality, differentiated product.
In FX, algorithmic trading is still a relatively new model, representing only about 15% of spot volume, but adoption is growing rapidly across both asset managers and corporates. Part of the reason for the increase in adoption is that asset managers and corporates are increasingly using trade cost analysis tools, which quantify the best execution benefits that algorithmic trading provides. In addition, the increasing fragmentation of liquidity across many venues and non-traditional trading counterparties is making it harder to achieve best execution by trading manually.
TM: Which market (equities or forex) is riper for innovation in algorithms? And why?
Mechner: We see innovation in both asset classes. Our perspective is as an algorithmic trading technology specialist supporting banks and brokers to develop a unique, custom algorithmic service for their customers. Larger banks and brokers are coming to us exactly because they need to innovate. In equities, brokers have heard from their customers that in a post-MIFID II world, “good enough” is no longer good enough. In FX, banks need to offer execution algorithms in order to stay relevant and top of mind with their clients, along with their other trading services.
In equities, even though algorithmic trading is mature, the pressure and level of innovation remains high – just across fewer providers. Not all providers of equity algorithms are able to make the continued investments in technology – for example in low-latency software, and in quantitative research to continually develop more powerful trading algorithms. In order to support our customers to stay at the front of the pack we are sourcing new tools and methodologies that have been developed at Google and Amazon but have not really penetrated capital markets.
The FX markets are overall still very much in an early adoption phase. Execution tools that have been commonplace in equities for ten years are still relatively novel but catching up quickly. Market participants have a large appetite for the analytics and analysis around execution algos, as they do in other asset classes, but in FX, as algo adoption is playing catch up, and the rate of incorporating analytics will be much faster. Additionally, as FX traders become more comfortable with algorithms, the number of customized algorithms will increase to better match customers’ trading objectives.
Across both asset classes, we see innovation in how our customers are designing their products to interact with the markets in unique ways, whether that’s incorporating unique liquidity pools, conditional orders, or how customers are able to express their trading goals.
TM: What are some of the latest developments in algos in equities?
Mechner: One major theme over the past couple of years is transparency.
The first driver of transparency is to protect customers and try to ensure that orders are being handled properly and in the customer’s best interests. This has been necessitated, in part, by many scandals over the past five years in which some of the largest banks and brokers were found to be handling customer orders differently than they advertised, to the detriment of their customers. Consuming the data that is the raw material of this transparency – for example detailed trade records – is creating increasing demand for more sophisticated and powerful analytics and data management tools and services.
The second driver of transparency is to provide actionable information that traders can use to do their job better – again, powered by increasingly powerful data management and visualization tools such as real-time analytics. Here, the market is still struggling to realize the marketing promise. The fundamental challenge is that financial data is complex, high-dimensional, noisy, and ultimately backward-looking. Distilling summary information that a human can consume and use to produce real-time decisions, but which can’t be consumed by algorithms in the first place, has proved challenging.
Another recent theme has been the marketing of algorithms using a new generation of machine learning tools, often dubbed “Artificial Intelligence.” This trend is driven by the stunning successes of AI tool sets over the past few years in domains from image recognition and natural language processing to chess and robotic control. However, here again in the algorithmic trading world, the marketing innovations may be leading the substance innovations. One challenge is that many firms were already using quantitative models and machine learning techniques without using the AI label. Another challenge is that now that many firms are using the AI label in their marketing, it doesn’t necessarily mean their clients can see a benefit. In the world of trading, sophisticated models don’t always translate to measurably enhanced execution performance. Though overall the market for execution tools has yet to see a clear benefit coming from the use of AI, Pragma is leveraging its AI expertise and actively investigating how these new AI tool sets can be used to bring measurable value to our clients.
TM: What are some of the latest developments in algos in FX?
Mechner: One relatively new development is that we’ve seen traders who have grown accustomed to the benefits of algorithms for Spot wanting to trade additional products with them, in particular NDFs. We launched an NDF algo in 2017, and continue to see increased adoption as the NDF market becomes more electronic.
TM: Are there other asset classes that Pragma is developing algos for such as fixed income, cryptocurrency or digital assets?
Mechner: We are always monitoring asset classes and regions that might present a compelling opportunity for our capabilities. However because each market has its own nuances and complexities, mastering a new market entails a significant, long-term investment, so we always weigh the value we can bring to our clients with new markets against improving our service in the markets we already cover.
While cryptocurrencies are certainly interesting, execution algorithms are fundamentally an institutional product, and we haven’t yet seen the necessary conditions – namely institutional demand – for crypto execution algorithms.
TM: Are these classes ripe for algo development? If so, why?
Mechner: The fixed income market is closest to achieving the structural conditions that allow execution algorithms to flourish. Like equities and FX, the market is becoming more electronic, more accessible to algorithmic trading, and more fragmented, particularly for U.S. Treasuries. The market is becoming more difficult for traders to trade by hand and thereby execution algos have the potential to be of value.
TM: Do regulators need to be better educated about the role of algorithms in financial markets?
Mechner: Regulators are generally well aware of the role of algorithmic trading in the financial markets. Over the past few years, we’ve seen a number of regulations across the globe focused on algorithmic trading covering everything from trading and monitoring, to software development lifecycle and testing, to change management.
One area we believe regulators would benefit from having a more in-depth familiarity with algorithmic trading is when considering market structure changes. We see a tendency for regulators to think in overly stylized terms about the roles of market participants. For example, regulators are used to thinking of the sell-side as price maker and the buy-side as price taker. But in truth, execution algorithms put institutions in competition with market makers to provide liquidity, and regulators’ analysis of market structure changes like the SEC’s proposed access fee pilot are sometimes not deep enough to reflect that reality when considering the impact their proposed changes will have on the market place.
Dealers have increased automation of manual electronic workflows.
Banks are competing for the top spot on investors’ algo wheels, Tabb report says.
Strategy looks to increase the targeted participation rate when the stock moves favorably.
Aim is to provide traders with optimal levels of transparency, performance and control.
A target-close model allows traders to mitigate market impact, Deutsche Bank's Andrew Royal writes.