The future of finance is a much-discussed topic, alongside the future of technology. The two, however, rarely meet outside the rarified environs of the Chief Technology or Chief Information Officer’s staff rooms and specialist conferences. Which is a pity, for the future of finance co-determines the future of technology, and the future of technology shapes the future of finance.
Over the next few posts, I will consider some of the technological innovation frontiers in financial services that are promising to dramatically alter the way financial markets function.
For starters, consider the favourite of the recent media and public attention: high frequency trading (HFT). In popular meme, this the focal point of ‘innovation gone wrong’.
However, in reality, HFT is a well-established sub-sector currently already mainstreamed into financial markets, but growing both in scale and scope.
On a positive side, it is a powerful co-player in the market alongside more traditional investors. For example, HFT technology can deliver improved efficiency of trading and distribution of information across larger markets in presence of information asymmetries.1HFT can also reduce market manipulation as well as lower trading costs, thus providing a significant liquidity benefit to the market.2 And despite the frequent popular arguments that HFT and algorithmic trading crowd out other investors, traditional investor still have capacity to succeed in the markets.3
In general, as algorithmic trading drives ever-greater reliance of investors on technological advantages of speed and machine-based information processing, the nature of financial markets is changing.
Not all of these are positive, and the constructive aspect of HFT is subject to numerous caveats, such as the environment in which markets operate or the nature of information flows in the markets.4
One recent study found that there is increasing concentration of excess returns within both the markets and the HFT sub-sector favours funds with aggressive, liquidity-sapping investment strategies. This reduces the availability of liquidity in the markets for passive investors, and concentrates markets returns in the hands of a small number of highly aggressive HFT funds, leading to the emergence of the ‘winner-takes-all’ markets.5
In part, advantages of technology-based trading strategies rely on pure speed of execution. But to an extent, they also rely on the speed and accuracy of automated data analytics built into these strategies. Beyond this lies the emerging ability of algorithms to ‘sense’ what is known as silent data or market signals that are simply not detectable from traditional data flows. For example, subtle, sub-order level changes in markets messaging traffic contained between the listed quotes have been shown to predict future price and quoted depth changes in the markets as much as 75 seconds in advance of these changes taking place. Modern electronic trading systems used by HFT firms can effectively forecast these short-term market conditions, where traditional investors generally fail to sense the changes in the market information.6
Technological change is transforming the markets in the directions that are hard to predict prior to their manifestations and increasingly difficult to regulate and manage. While some of these transformations benefit the markets, others create harmful disruptions and can alter the very nature of the market itself, impairing longer-term objective of supplying real investment to the real economy.
At the technological bounds of trading, today, two key problems raise serious grounds for concern.
The first one is the uncertain nature of market structures as the speed of transactions engaged in by the HFT algorithms rises to its natural limit. In physics, matter behaves in categorically distinct ways as it approaches the speed of light. So do information – the oxygen of the asset markets. Investors, located at different physical or market points (such as geographically separated investors or investors transacting in separated, but related markets, such as markets for equities and derivatives) may simultaneously witness different “best” prices quoted. Meanwhile, regulators and markets systems may not be able to distinguish the timing of individual orders and transactions. This can result in the situation where best execution principles can no longer be upheld and market rules and regulations can become ‘fungible’.
The second area of concern is that of the extent of information capture by the machine-readable analytics software. Artificial intelligence, higher complexity mathematical modelling tools, faster processing speeds of supercomputers and nascent, but growing ability of language-based software systems to analyse qualitative data in machine-readable formats all are actively displacing traditional human analysis and intuition.7
These factors interact with the very nature of the physical technology used in finance today. Supercomputers exhibit rapid rates of depreciation in terms of returns on investment.8 Which means that technologically advanced trading firms tend to be bigger, more institutionalised, squeezing the smaller players. These, and other aspects of the technological transformation ongoing in the financial services are putting new and ever-rising pressures on our ability to manage systemic risks, and once again turning financial services sector into a race between investors, supervisors, regulators and markets providers for comparative technological advantage.
As one of the researchers cited above noted: “Modern finance is becoming cyborg finance — an industry that is faster, larger, more complex, more global, more interconnected, and less human.”9
Dr Constantin Gurdgiev is the Adjunct Assistant Professor of Finance with Trinity College, Dublin
References
1. Huh, Yesol, Machines vs. Machines: High Frequency Trading and Hard Information (March 4, 2014). FEDS Working Paper No. 2014-33
2. Aitken, Michael J. and Aspris, Angelo and Foley, Sean and Harris, Frederick H. deB., The Effects of Algorithmic Trading on Security Market Quality (April 30, 2014).
3. Easley, David and Lopez de Prado, Marcos and O’Hara, Maureen, The Volume Clock: Insights into the High Frequency Paradigm, The Journal of portfolio management, Fall 2012, Vol. 39, No. 1: pp. 19-29
4. Brogaard, Jonathan and Hendershott, Terrence and Riordan, Ryan, High Frequency Trading and the 2008 Short Sale Ban (October 13, 2014), Foucault, Thierry and Kozhan, Roman and Tham, Wing Wah, Toxic Arbitrage (April 2014). CEPR Discussion Paper No. DP9925. and Menkveld, Albert J. and Zoican, Marius A., Need for Speed? Exchange Latency and Liquidity (July 22, 2014). Tinbergen Institute Discussion Paper 14-097/IV/DSF78.
5. Baron, Matthew and Brogaard, Jonathan and Kirilenko, Andrei A., Risk and Return in High Frequency Trading (May 5, 2014).
6. Harris, Jeffrey H. and Saad, Mohsen, The Sound of Silence (May 2014). Financial Review, Vol. 49, Issue 2, pp. 203-230, 2014.
7. Gurdgiev, Constantin and Saxton, Keith, Data: At the Core of the Problem (April 25, 2012). Journal of Central Banking, pp. 87-94, March 2012. and Lin, Tom C. W., The New Financial Industry (March 30, 2014). 65 Alabama Law Review 567 (2014); Temple University Legal Studies Research Paper No. 2014-11.
8. Apon, Amy and Ahalt , Stanley and Dantuluri, Vijay and Gurdgiev, Constantin and Limayem, Moez and Ngo, Linh and Stealey , Michael, High Performance Computing Instrumentation and Research Productivity in U.S. Universities (September 19, 2010). Journal of Information Technology Impact, Vol. 10, No. 2, pp. 87-98, 2010.
9. Lin, 2014 cited in note 7 above
Originally posted at http://blog.learnsignal.com/?p=111
