Please use this identifier to cite or link to this item:
Title: A complex systems approach to financial markets, social learning and socioeconomic phenomena
Authors: Loh, Lyon Han Zhou
Keywords: DRNTU::Social sciences::Economic theory
Issue Date: 2017
Source: Loh, L. H. Z. (2017). A complex systems approach to financial markets, social learning and socioeconomic phenomena. Doctoral thesis, Nanyang Technological University, Singapore.
Abstract: This thesis studies complexity in financial markets, social learning and socioeconomic phenomena. Economies and societies are complex systems, with decentralization and local interaction of numerous people with heterogeneous behavior. The thesis departs from the independent rational agent paradigm and adopts the generativist approach of how a population of bounded rational and heterogeneous agents, through micro-interactions, can emerge to a different kind of macro-behavior of the system. Chapter 1 describes the research background/approach of this thesis and relates complexity to the history of economic thought. Chapter 2 explicates how the salient characteristics of financial prices emanates from mass psychology and investor sentiments explicitly by modeling the interactional dynamics between herding and contrarian behavior of investors in financial markets using a computational agent-based approach. Utilizing a cellular automaton agent-based model with heterogeneous agents, herding and contrarian investors make their investment decisions based on interactions with other investors in a regular network. The agents in our model determine their investment decisions based on the actions of other agents in their network neighborhood which in turn are based on the actions of others in their respective network neighborhoods. The model succeeds in providing a generative explanation of rich and complex characteristics of financial markets such as chaotic price fluctuations, punctuated equilibria, fat tails, extreme values, market cycles, non-periodic bubbles and crashes, and switching regimes of bull and bear markets. Chapter 3 utilizes an agent-based model to explore the non-pecuniary impact of social learning on the utilitarian efficiency of the population in decision-making between discrete choices. The model embeds a small-world network into a cellular automaton framework. Consumers augment their limited private information by learning and deriving information from social interaction and imitation in their social network. The social learning between individual consumers in a small-world network is an efficient aggregation mechanism of information which overcomes the frictions of imperfect information and bounded rationality of individuals. Social learning in liberal economies generates aggregate rationality from individual bounded rationality, resulting in the emergence of self-organized efficiency of consumers in optimal decision-making. Chapter 4 explores the dynamics of socio-economic phenomena using a complex systems approach. Using a stochastic cellular automata network model (CANM), bounded rational agents herd to the actions of other agents in their social network, a convergent social behavior resultant of social influence. The CANM model allows for an interdependent multiple-agent paradigm which structures the local interaction of agents in a small-world network. This paper offers a generative explication of various socio-economic phenomena as the emergence of Quasi-herding interaction among individuals, and in addition relates the results to examples not restricted to skirts, cohabitation, political revolutions and individualism. A parsimonious model of quasi-herding behavior elucidates outcomes of fleeting fads, (cyclical) fashion, non-periodic cycles of bubbles and crashes, changing regimes of social norms, and lock-in of cultures. Chapter 5 summarizes and concludes the thesis.  
DOI: 10.32657/10356/71650
Fulltext Permission: open
Fulltext Availability: With Fulltext
Appears in Collections:HSS Theses

Files in This Item:
File Description SizeFormat 
Thesis_Lyon.pdfThesis2.96 MBAdobe PDFThumbnail

Page view(s) 50

Updated on May 16, 2021

Download(s) 50

Updated on May 16, 2021

Google ScholarTM




Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.