The Global Financial Markets

The LSCITS Initiative initiated and helped to lead a major international study on markets that could have global impacts.

FOR ALMOST HALF a century, since the early 1970s at least, traders in the global financial markets have been using computer terminals of increasing sophistication and complexity. In the last decade in particular, there has been a rapid growth in the development of automated trading systems that replace human traders, and that can in principle trade large numbers of different assets across multiple trading venues, on multiple continents, integrating vast quantities of data and reacting at superhuman speed.

The global financial markets are now clearly a planetary-scale interconnected sociotechnical supersystem and a perfect example of a large-scale complex IT system. As such, there are concerns that it may be poorly understood, and may be liable to types of potential failure that are only made apparent when those failures actually occur - a prospect that, given the manifest economic signature of the markets, could have devastating consequences for individual investors, companies, or even entire national or transnational economies.

The LSCITS Initiative's work studying the global financial markets has been conducted mainly by a team at the University of Bristol. In late 2009, Professor Dave Cliff was requeted by the UK Government's Chief Scientific Advisor, Professor Sir John Beddington, to write a brief proposal for a possible study on the future implications arising from the way in which the financial markets had become large-scale complex IT systems.

The final version of that proposal was submitted to the UK Government Office for Science (GOS) in April 2010, and less than two weeks later a major malfuntion occurred in the US financial markets, when prices gyrated wildly on the afternoon of 6th May 2010, shedding and regaining almost $1 trillion in one hour of turmoil. Shortly after this so-called Flash Crash event, the GOS announced that its Foresight unit would run a two-year intensive study on the future of computer trading in the financial markets (FCTFM).

The Foresight FCTFM project commissioned survey papers and original research from 150 leading academics around the world, and all of the project's publications were subject to anonymous peer review by two or more internationally leading academics prior to publication under Crown Copyright on the GOS website. Members of the Bristol LSCITS Initiative team were involved in authoring or co-authoring six major papers published through the Foresight project, more than 10 per cent of the projects total output of 53 such publications.

Cliff co-authored a 22,000-word position paper with Linda Northrop -SEI Fellow and Director of the Ultra-Large-Scale Systems Program at the Software Engineering Institute, Carnegie Mellon University, USA - and also a 15,000-word survey paper with Dr Dan Brown and Professor Philip Treleaven, of the UK Financial Computing Doctoral Training Centre based as University College London. Bristol LSCITS Initiative postdoctoral fellow Dr John Cartlidge and PhD students Marco de Luca and Charlotte Szostek worked together on two Foresight papers that reported on experiments studying human traders interacting with algorithmic trading systems, and the source code for these experiments has been released under an open-source licence, as a free resource for other researchers around the world. Building on this work, Bristol LSCITS Initiative undergraduate intern Steve Stotter's final-year research project studied interactions among automated trading algorithms, and was awarded best paper prize at an international conference held in Italy.