Markets were primarily designed to help buyers and sellers discover prices and volumes of traded items. But, market deregulation, technological innovation and ideological changes, have led to an epic financialisation of stock and commodity trading. These developments have led to frequent booms and busts both at the local level and sometimes globally as well. With our age of nudging, framing, and fungibility, psychologists and computer professionals are creating market models where knowledge based on experience, codified as rules of thumb, outperforms sophisticated econometric models, and we have algorithmic trading where the reactions of the adversaries are anticipated like moves on a chess board. Decisions based on emotional intelligence, that is on working out the sentiment – hopes and fears - of others, appears to supplant decisions based on rational behaviour, so say the high priests of behavioural finance –including Nobel Laureates Daniel Kahneman, Robert Shiller, and Richard Thaler, inspired in different ways by Benoit Mandelbrot. Artificial intelligence systems now exist that can detect emotional language in reports and comments, and can detect leakage of ‘true’ emotions in voice and facial expressions. These sentiments are quantified and used in conjunction with econometric models. The hybrid behavioural finance models have a 5-10 basis points advantage in stock and indices trading, and up to 10-15 basis advantage in commodity trading. These AI-based models do move the markets in laboratories, can these models survive in the hustle and bustle of the (virtual) trading floor? That is the question for you to ask and for me to speculate.
Professor Khurshid Ahmad is the Professor of Computer Science in the School of Statistics and Computer Science, Trinity College Dublin. His research areas include artificial intelligence, neural networks, fuzzy logic, social media analytics and behavioural finance. He was trained as a nuclear physicist and has worked in high-performance computing covering areas such as forecasting, computer-assisted learning, engineering design, and information extraction from continuous information streams comprising texts, images and numbers. His work seeks to maximise the potential of computing systems by enabling these systems to deal with different modalities of human communications, language, vision, symbolic including numerical information exchange. He has designed and implemented systems that learn to deal with the different modalities of communications. His work has been supported by research councils, EU Programmes, and venture capital funds. He is a former Visiting Professor at Copenhagen Business School and the University of Surrey, and has worked with UN FAO and UNDP. He has published over 200 research papers and his work has appeared in journals in AI and in corporate finance. His latest book is on the topic of Social Computing and the Law (Cambridge University Press). He is a Fellow of the British Computer Society and of Trinity College, Dublin. He is a member of the Board of Trinity College Dublin.