According to a recent Wall Street & Technology article titled "Where is the Innovation in Trading Technology?" participants of a recent TABBForum Conference have complained about the lack of innovation in trading technologies. Some of them going as far as claiming "massive innovation gaps" and attesting to the existence of "workflow inefficiencies preventing a trader from covering an account."
Notwithstanding the innovations in latency reduction and hardware and middleware environment improvement over the first decade of the 21st century, a gap of innovation indeed can be observed in trading technologies.
Much of this gap should be attributed to one of the least discussed subjects -- trading decision making and management analytics. The subject is sometimes considered the Holy Grail of trading analytics, and is frequently surrounded by a shroud of mystery -- possibly due to the need for technical and data skills in order to assess the true state of the discipline.
It is also possible that some in the industry think that since a lot of money has been spent on the development of trading analytics and armies of highly qualified people have worked on it, that all of the meaningful work has already been completed.
Meanwhile, detached from the subject of analytics is the buzz surrounding algorithmic and high-frequency trading (HFT). Much of the buzz associated with HFT is whether it is good or bad for the overall markets and if HFT (as a subset of algorithmic trading) should be left to its own devices and market forces, or be curtailed and restricted through regulatory and legislative activism.
How Do You Use Your Algo?
More often than not, algorithmic trading is presented as the substance rather than a method of injecting operational efficiency, productivity and optimization via the automation of quantified sequences of actions, supposedly leading to some well-conceptualized and well-defined goal. Algorithmic trading is not what one does but how one does it.
Furthermore, ever since HFT gained in popularity, the subject has been trivialized by reducing it to latency measurements -- which gives a false impression that all one needs to do is execute a trade faster and that all the misfortunes of the other guy are resulted from his lack of access to fast execution capabilities.
So before immersing further into the trading technology innovation debate, one needs to also consider the overall climate surrounding the discipline of trading, portfolio and asset management.
Outside of the financial industry, especially in some political and politicized circles, a question is also asked as why one needs algorithmic trading. Essentially, the question is whether trading constitutes leisure for people with high degree of passion for gambling or it is a needed mechanism for supporting the base economy?
The Trading "Game"
The short answer to this question is that trading is a game, however not in the sense of a leisure activity. It is a mathematical game of quantification of the behavioral dynamics of market participants and manifests a strategic competition for the limited risk-transfer and risk-monetization resources of financial markets.
Trading overall is a risk and risk-premium pricing mechanism serving the management of uncertainty -- an inherent and unalienable attribute of financial markets. It enables corporations and financial intermediaries to transfer and monetize risk and raise capital for the benefit of the base economy.
As algorithmic trading serves the purpose of making the markets more efficient and execution more productive, one can't help but to think that the attempts of curtailing or restricting algorithmisation of trading is similar to the protests of coachmen directed at the automobile industry at the beginning of the 20th century. It is also noteworthy that the transition of coachmen to automobile drivers was not an automatic one since it was more likely that more readily the mechanics would get driving jobs than the coachmen.
The material background on which algorithmisation of trading is taking place today is the shrinking of trading revenues and volumes with the increasing operational costs, including those of IT. Furthermore liquidity is diminished due to the proliferation of trading venues and increased risk-aversion of market participants. All of the above make the financial markets' dynamics more complex and diminish the traditional hidden value monetization sources. And there is a belief in the advanced quarters of the financial industry that the current situation is the "new normal" rather than a transient crisis.
Under these circumstances financial securities industry doesn't have any other choice but to digitize every remaining quarter and process, automate, rationalize and optimize, and this natural course of financial markets' process optimization and efficiency and productivity improvement cannot be reversed unless through a transition to a full- or semi-totalitarian system of regulation and governance. And while one can wish to claim that history has given us enough and rather picturesque examples of the consequences of autocratic, semi- or fully-totalitarian management of complex systems, perhaps some more examples of such consequences are yet to be witnessed in the coming years.
Meantime, based on what has transpired over the last few years in the financial trading, portfolio and asset management industry, a few things need to be viewed in their proper perspective.
A Cure All?
For instance, when one hears conversations of a trading algorithm being better than that of the another firm, or of an algorithm being "the cure of all ills," or in much the same way, the algorithms being the "source of all ills," a few important questions should be asked: What problem does the given algorithm solve? Why does it solve it? And, why is it the best or optimal solution of that problem?
Due to the fundamental uncertainty embedded in the future dynamics of financial markets and their non-stationary character, no historic performance can serve as justification of a given algorithm.
Justification of any algorithm is to be found in its underlying model's accurate replication of the corresponding business processes and in the sufficiently realistic capturing of the financial market dynamics through the utilization of contextually relevant and well-defined quantitative metrics for financial market participants' risk perception and risk-premium.
Assuming that the justification is found, the next question to be asked is whether the algorithmic activity -- both on decision making and execution levels -- is supported by an adequate software, platform and infrastructure guaranteeing timely, efficient and reliable execution and adequate pre- and post-execution quality assurance.
As one can hear conversations on about 500 or 600 pre-set trading algorithms drifting in the public domain or within the secretive quarters of hedge funds and proprietary trading houses, it is important to note that having preset algorithms which allegedly perform as desired in some pre-set situations does not present an impressive treasure. Nor it should serve as a deep pretext for a meaningful conversation as all algorithms should be subjected to a "lucky or smart" scrutiny.
Do No Harm
Finally, algorithmic trading can be compared to playing jazz in jam-sessions with musicians one might have never met before. For instance, to play well in jazz the participants need to have performance confidence induced by a flawless technique, know the theme and harmony, have the ability to recognize on-the-fly stage-situation and react relevantly quickly, non-abrasively and with due respect to the rest of the players.
With a clear understanding that they provide a crucial service to the financial markets, innovations in algorithmic trading and automation of all relevant processes need to pursue the mission of supporting the ecosystem they operate in instead of disrupting it.
Streamlining, automation, optimization and commoditization of trading decision making and management processes for the trading, portfolio and asset management lifecycles will be the next driver and will open the floodgates of innovation in trading technologies.
For sure, decision making automation will define how software, platform and infrastructure will need to be innovated in the coming future. This next level of innovation, along with systematization and optimization of the trading industry, will require further advancement of trading analytics, elimination of some old controversies, descriptive metrics for the measurement of the expected behavior of portfolios, supposedly resulting from the submission of trade orders.
Finally, quantification and management of the diminished liquidity and its factoring into decision making analytics will be one of the fundamental components of the next generation trading systems.
About The Author: Marti Jermakyan, the CEO of RACS, LLC, a resident of the Skolkovo Innovation Center in Moscow, RF, is a recognized expert in financial and commodity markets quantitative trading methods and applications with a long academic and senior level entrepreneurial and corporate history. He has started his financial markets career in Chicago in the era of the beginning of the electronification of financial markets and working with the pioneers of this industry.The InformationWeek community brings together IT practitioners and industry experts with IT advice, education, and opinions. We strive to highlight technology executives and subject matter experts and use their knowledge and experiences to help our audience of IT ... View Full Bio