Principal Investigator: Professor Nick Chater
Professor Nick Chater is Professor of Behavioural Science at Warwick Business School, having previously held chairs in Psychology at Warwick and UCL. He is particularly interested in the cognitive and social foundations of rationality, and in applying behavioural insights to public policy and business. He has served as Associate Editor for the journals Cognitive Science, Psychological Review, Psychological Science and Management Science. He co-founded, and is a Director of, the research consultancy Decision Technology Ltd, and is a member of the Committee for Climate Change. He was elected a Fellow of the British Academy in 2012.
Recent developments in cognitive science have created new models of the moment by moment calculations by which people make decisions. How far might these models also help to explain puzzling economic phenomena? This project aims to apply cutting-edge computational models of cognition to help explain both ubiquitous “scale-invariant” patterns in financial time series, such as stock market prices, and also to connect with existing psychologically-based economic models of booms and crashes.
Adam Sanborn (Department of Psychology, University of Warwick) and Nick Chater (Warwick Business School), working with PhD student Zhu Jianqiao, have been exploring the idea that people deal with uncertainty by approximating the enormously complex calculations that would be recommended by probability theory (Sanborn & Chater, 2016; Zhu, Sanborn & Chater, 2017, in prep)—the Bayesian ‘rational standard’ that is usually the starting point for research in psychology and economics. A particularly appealing class of approximations operate by sampling: choosing particular instances from a probability distribution and drawing conclusions from those instances. With an infinitely large sample, this approach would converge on Bayesian rationality. But, in reality, it is more likely that the brain works with very small samples and thus will generate many biases and errors compared to the rational standard.
Their research has already shown that many classic probabilistic reasoning errors can be explained as arising from the sampling process. Moreover, a specific approach to sampling developed within computational statistics, and known as MC3, turns out accurately to match ‘scale-free’ patterns in human behaviour, explaining variation within trials, and over time, for individual experimental participants.
Interestingly, financial time series (e.g., daily return index such as Dow Jones Composite and S&P500) also displays fat-tails and long-range dependencies (Mandelbrot, 2013), which are mathematically closely related to the patterns found in individual human behaviour. It is possible that these scale-free properties are the result of the aggregation of many individual actors performing approximate inference, but with biases that may add rather than cancel each other out.
Moreover, it is possible that the approximate nature of sampling at the level of the individual decision-maker may lead to systematic biases which can leave the entire market into unexpected and non-rational behaviour. Andrei Schleifer and colleagues (e.g., Gennaioli & Shleifer, 2018) have recently argued that a particular cognitive bias, known as representativeness (roughly, an excessive focus on stereotypical scenarios, Tversky & Kahneman, 1974), leads markets to overreact in the light of new information positive or negative; and that such overreactions can lead to market instability, including booms and crashes. A second element of our project laws how limitations on sampling may result in the psychological bias of representativeness, and potentially provide a new microfoundation for Scheifer’s analysis.
Gennaioli, N., & Shleifer, A. (2018). A Crisis of Beliefs: Investor Psychology and Financial Fragility. Princeton University Press.
Mandelbrot, B. B. (2013). Fractals and scaling in finance: Discontinuity, concentration, risk. Selecta volume E. Springer Science & Business Media.
Sanborn, A. N., & Chater, N. (2016). Bayesian brains without probabilities. Trends in Cognitive Sciences, 20(12), 883-893.
Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases. Science, 185(4157), 1124-1131.
Zhu, J. Q., Sanborn, A. N., & Chater, N. (2017). Mental Sampling in Multimodal Representations. arXiv preprint arXiv:1710.05219.
Zhu, J-Q., Sanborn, A. N., & Chater, N. (in prep). More rational: sampling plus correction explains biases in judgments.
Results will be published here when available.