Purpose: The aim of the present work is to design a model of stock market composed by virtual investors and to consider individual differences among these. Design/Methodology: For this purpose we made use of an Agent-Based Model (ABM), where each agent in the simulation represent an individual investor capable to move in the virtual environment and to make transactions with other agents. Each agent was initially fitted with a virtual portfolio of investments and was generated comprising: (1) a personal risk tolerance, modelled within a loss-aversion perspective (2) time-life (3) an investment objective (4) a learning behaviour based on experience. The former and the latter features were based on a previous paper (Ceschi, Rubaltelli e Sartori, 2014) which implemented individual differences into the Value function (Kahneman, Tversky, 1979). Results: Different scenarios were simulated changing the learning behaviour criteria and the number of investors within the simulation. All different scenarios generated are discussed in a comparison perspective. Limitations: A limitation of the present study lies in the number of variables computed into the simulation model. Research/Practical Implications: Nevertheless, the simulation is able to forecast different scenarios of stock markets considering individual differences (namely, the sensitivity to losses) among the investors. Originality/Value: The present study is extremely valuable since it allows exploring many possible scenarios disclosing relevant consequences that can arise from psychological tendencies induced by positive and negative investment performance.
Using an Agent-Based Model to Simulate Loss-aversion and Learning Behaviour among Investors
Scalco, Andrea;CESCHI, Andrea;
2015-01-01
Abstract
Purpose: The aim of the present work is to design a model of stock market composed by virtual investors and to consider individual differences among these. Design/Methodology: For this purpose we made use of an Agent-Based Model (ABM), where each agent in the simulation represent an individual investor capable to move in the virtual environment and to make transactions with other agents. Each agent was initially fitted with a virtual portfolio of investments and was generated comprising: (1) a personal risk tolerance, modelled within a loss-aversion perspective (2) time-life (3) an investment objective (4) a learning behaviour based on experience. The former and the latter features were based on a previous paper (Ceschi, Rubaltelli e Sartori, 2014) which implemented individual differences into the Value function (Kahneman, Tversky, 1979). Results: Different scenarios were simulated changing the learning behaviour criteria and the number of investors within the simulation. All different scenarios generated are discussed in a comparison perspective. Limitations: A limitation of the present study lies in the number of variables computed into the simulation model. Research/Practical Implications: Nevertheless, the simulation is able to forecast different scenarios of stock markets considering individual differences (namely, the sensitivity to losses) among the investors. Originality/Value: The present study is extremely valuable since it allows exploring many possible scenarios disclosing relevant consequences that can arise from psychological tendencies induced by positive and negative investment performance.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
