Development of an agent-based speculation game for higher reproducibility of financial stylized facts

Abstract

Simultaneous reproduction of all financial stylized facts is so difficult that most existing stochastic process-based and agent-based models are unable to achieve the goal. In this study, by extending the decision-making structure of Minority Game, we propose a novel agent-based model called "Speculation Game," for a better reproducibility of the stylized facts. The new model has three distinct characteristics comparing with preceding agent-based adaptive models for the financial market: the enabling of nonuniform holding and idling periods, the inclusion of magnitude information of price change in history, and the implementation of a cognitive world for the evaluation of investment strategies with capital gains and losses. With these features, Speculation Game succeeds in reproducing 10 out of the currently well studied 11 stylized facts under a single parameter setting. nancial time series [1]. The 11 currently reported stylized facts can be listed as follows: volatility clustering, intermittency, heavy tails, the absence of autocorrelation in returns, slow decay of autocorrelation in volatilities, volume/volatility correlation, aggregational Gaussianity, conditional heavy tails, asymmetry in time scales, leverage effect and gain/loss asymmetry. They are quite nontrivial features which have been well studied for many years in different markets and for various instruments.The reproducibility of the stylized facts is a prerequisite for the financial market model. However, to reproduce them with a conventional market model, whether a simple or sophisticated stochastic process, is not very easy due to the high level of multiplexity of the market. With the contribution of econometrists, several well known stochastic models have been constructed to yield typical stylized features, for example, the autoregressive conditional heteroscedasticity (ARCH) and generalized autoregressive conditional heteroscedasticity (GARCH) processes [2,3]. Nonetheless, to the best of our knowledge, most existing models fail to reproduce the whole set of stylized properties simultaneously.On the other hand, an alternative approach mainly advocated by econophysicists, which adopts agent simulation based on a set of simple rules, has been used to study the stylized facts as well. The simulation enables the reproduction and analysis of complex price dynamics in an artificial market with autonomous agents acting as the traders. Compared with previous stochastic process models, which are designed to describe the traits of financial time series at the macroscopic level, this method is a bottom-up approach which explains these features at the microscopic level, that is, the level at which traders' decision-making and trading behaviors are taking into account.There are mainly two reasons for adopting agent-based models. First, the agent-based simulation is almost the only way to analyze the effects of traders' behavior to the whole market. All time-series data obtained from the real market, such as price and volume, are aggregational ones, which i...

On the Relationship between Investment Risk and Insurance Solvency Optimum

“…Although asset return distributions in practice may deviate from normal distribution (e.g. Albuquerque, 2012;Katahira et al, 2019), in the model it can be assumed that the ρ investment return follows normal distribution, and the investment risk may be measured by the standard deviation of the random variable ρ. As classical financial theory literature suggests (e.g.…”

Heterogeneous wealth distribution, round-trip trading and the emergence of volatility clustering in Speculation Game

“…Based on these previous studies, we developed Speculation Game [29], which is an adaptive agent model allowing both an explicit description of round-trip trading and a dynamic evolution of traders' wealth. In Speculation Game, round-trip trading is defined as a trading process in which an agent, responding to a set of signals received from the price movements, opens a position with a buy/sell order first, and closes the position with the reverse order later.…”