Abstract
I introduced a good AI for finance, an artificial market (an agent-based model for a financial market) to design a financial market that works well, and a bad AI for finance, an AI trader that discovers how to manipulate markets through learning in an artificial market. First, I describe an artificial market for designing a financial market that works well. Artificial markets have recently been used to develop and examine rules and regulations such as tick size reduction in actual financial markets. Their influence is growing. A good market design is important for developing and maintaining an advanced economy. Second, I show that an AI trader that learns in an artificial market can discover how to manipulate a market and use that capability as an optimal investment strategy even when the developer of the AI has no such intention, and the result suggests the need for regulation. These stories show that there are both good and bad AIs for society. We should not discuss whether AIs on the whole are good or bad, but rather how to manage them so that they are useful tools. To do that, we should discuss rules, ethics, and philosophy for AIs.
Note that the opinions contained herein are solely those of the authors and do not necessarily reflect those of SPARX Asset Management Co., Ltd.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
Notes
- 1.
Sabzian et al. provide a comprehensive review of agent-based models for complex systems (Sabzian et al. 2018).
- 2.
A stylized fact is a term used in economics to refer to empirical findings that are so consistent (for example, across a wide range of instruments, markets, and time periods) that they are accepted as truth (Sewell 2011).
- 3.
- 4.
This enables us to focus on phenomena on short time scales, as the fundamental price remains static.
- 5.
When t < τ j, however, I set the second term in Eq. (13.1) to 0.
- 6.
When t < t c, however, to generate enough waiting orders, the agent places an order to buy one share when \(P_f>P^t_{o,j}\) or to sell one share when \(P_f<P^t_{o,j}\).
- 7.
Darley and Outkin (2007) investigated tick size reduction using an artificial market model when NASDAQ, a stock exchange in the USA, was planning a tick size reduction. They showed, for example, that a market can become unstable when some investment strategies become more prevalent. However, their model had too many parameters and they focused on which investors earned more, which prevented them from examining factors contributing good design of financial markets.
- 8.
https://ai-finance.org/. It was originally scheduled to be held in New York but was held virtually.
- 9.
I presented there together with co-author Yagi et al. (2020).
- 10.
A genetic algorithm is a calculation method inspired by evolution and natural selection that searches for an approximate optimal solution. Input values are represented as genes, and a surviving gene that has a higher adaptability (output value) leads to an optimal solution, that is, the input value that results in the highest output value (Goldberg 1989).
- 11.
In reality, the builder always intends certain strategies to be selected when modeling possible strategies. However, it is crucial for this study that the developer has no intention of any strategies including market manipulation. Therefore, I did not intentionally model trading strategies, and my model directly searched for all the optimal trades in an artificial market environment. Because there are no models of trading strategies, my model does not output in an out-sample forecast; thus, no one can test my model in an out-sample forecast. I argue, however, that this study does not need such evaluations because it focuses on whether an AI trader can discover market manipulation strategy through its own learning despite the builder having no intention of market manipulation. This study does not aim to use the model to out-sample forecast.
- 12.
AIA does not take any action before tick time t c to stabilize the simulations. As mentioned in footnote 7, the purpose of the period before t c is to generate enough waiting orders.
References
Arthur W, Durlauf S, Lane D, Program SE (1997) Asset pricing under endogenous expectations in an artificial stock market. The economy as an evolving complex system II, pp 15–44
Battiston S, Farmer JD, Flache A, Garlaschelli D, Haldane, AG, Heesterbeek H, Hommes C, Jaeger C, May R, Scheffer M (2016) Complexity theory and financial regulation. Science 351(6275):818–819. https://doi.org/10.1126/science.aad0299
Bookstaber R (2017) The end of theory: Financial crises, the failure of economics, and the sweep of human interaction. Princeton University Press
Borges JL (1954) Del rigor en la ciencia. In: Historia universal de la infamia. Emecé
Braun-Munzinger K, Liu Z, Turrell A (2016) Staff working paper no. 592 an agent-based model of dynamics in corporate bond trading. Bank of England, Staff Working Papers. https://www.bankofengland.co.uk/working-paper/2016/an-agent-based-model-of-dynamics-in-corporate-bond-trading
Byrd D, Hybinette M, Balch TH (2019) Abides: Towards high-fidelity market simulation for ai research. https://arxiv.org/abs/1904.12066
Chakraborti A, Toke IM, Patriarca M, Abergel F (2011) Econophysics review: II. Agent-based models. Quantitative Finance 11(7):1013–1041. https://doi.org/10.1080/14697688.2010.539249
Chen SH, Chang CL, Du YR (2012) Agent-based economic models and econometrics. Knowl Eng Rev 27(2):187–219. https://doi.org/10.1017/S0269888912000136
Chiarella C, Iori G (2002) A simulation analysis of the microstructure of double auction markets. Quantitative Finance 2(5):346–353. https://doi.org/10.1088/1469-7688/2/5/303
Collver C (2017) An application of agent-based modeling to market structure policy: the case of the U.S. tick size pilot program and market maker profitability. In: White Paper. U.S. Securities and Exchange Commission. https://www.sec.gov/marketstructure/research/increasing-the-mpi-combined.pdf
Cristelli M (2014) Critical review of agent-based models, pp 29–54. Springer International Publishing, Cham. https://doi.org/10.1007/978-3-319-00723-6_3
Darley V, Outkin AV (2007) Nasdaq market simulation: Insights on a major market from the science of complex adaptive systems. World Scientific Publishing
Farmer JD, Foley D (2009) The economy needs agent-based modelling. Nature 460(7256):685–686. https://doi.org/10.1038/460685a
Foucault T, Menkveld AJ (2008) Competition for order flow and smart order routing systems. J Finance 63(1):119–158. https://doi.org/10.1111/j.1540-6261.2008.01312.x
Goldberg DE (1989) Genetic algorithms in search, optimization, and machine learning. Addison-Wesley Professional
Haruvy E, Noussair CN (2006) The effect of short selling on bubbles and crashes in experimental spot asset markets. J Finance 61(3):1119–1157. https://doi.org/10.1111/j.1540-6261.2006.00868.x
Hommes C, Breen K (2018) Integrated macro-financial modelling for robust policy design. https://cordis.europa.eu/project/id/612796/reporting
Izumi K, Okatsu T (1996) An artificial market analysis of exchange rate dynamics. Evolutionary programming V, proceedings of the fifth annual conference on evolutionary programming, pp 27–36
Izumi K, Toriumi F, Matsui H (2009) Evaluation of automated-trading strategies using an artificial market. Neurocomputing 72(16):3469–3476. https://doi.org/10.1016/j.neucom.2008.07.020. Financial Engineering Computational and Ambient Intelligence (IWANN 2007)
Karpe M, Fang J, Ma Z, Wang, C (2020) Multi-agent reinforcement learning in a realistic limit order book market simulation. https://arxiv.org/abs/2006.05574
Kita H, Taniguchi K, Nakajima Y (2016) Realistic simulation of financial markets. Springer. https://doi.org/10.1007/978-4-431-55057-0
Kuo CH, Chen CT, Lin SJ, Huang SH (2021) Improving generalization in reinforcement learning based trading by using a generative adversarial market model. IEEE Access 9:50738–50754. https://doi.org/10.1109/ACCESS.2021.3068269
LeBaron B (2006) Chapter 24 agent-based computational finance. Elsevier, pp 1187–1233. https://doi.org/10.1016/S1574-0021(05)02024-1
Lux T, Marchesi M (1999) Scaling and criticality in a stochastic multi-agent model of a financial market. Nature 397(February):498–500. https://doi.org/10.1038/17290
McMillan J (2002) Reinventing the bazaar: A natural history of markets. W. W. Norton & Company
Menkhoff L, Taylor MP (2007) The obstinate passion of foreign exchange professionals: technical analysis. J Econ Lit, 936–972. https://doi.org/10.1257/jel.45.4.936
Mitchell M (2009) Complexity: A guided tour. Oxford University Press
Mizuta T (2016) A brief review of recent artificial market simulation (agent-based model) studies for financial market regulations and/or rules. SSRN Working Paper Series. https://doi.org/10.2139/ssrn.2710495
Mizuta T (2020) An agent-based model for designing a financial market that works well. In: 2020 IEEE symposium series on computational intelligence (SSCI), pp 400–406. https://doi.org/10.1109/SSCI47803.2020.9308376
Mizuta T (2020) Can an ai perform market manipulation at its own discretion?-a genetic algorithm learns in an artificial market simulation-. In: 2020 IEEE symposium series on computational intelligence (SSCI), pp 407–412. https://doi.org/10.1109/SSCI47803.2020.9308349
Mizuta T, Hayakawa S, Izumi K, Yoshimura S (2013) Investigation of relationship between tick size and trading volume of markets using artificial market simulations. In: JPX working paper, 2. Japan Exchange Group. https://www.jpx.co.jp/english/corporate/research-study/working-paper/index.html
Mizuta T, Kosugi S, Kusumoto T, Matsumoto W, Izumi K, Yagi I, Yoshimura S (2016) Effects of price regulations and dark pools on financial market stability: An investigation by multiagent simulations. Intell Syst Account Finance Manag 23(1–2):97–120. https://doi.org/10.1002/isaf.1374
Nagumo S, Shimada T, Yoshioka N, Ito, N (2017) The effect of tick size on trading volume share in two competing stock markets. J Phys Soc Jpn 86(1):014801. https://doi.org/10.7566/JPSJ.86.014801
O’Hara M, Ye M (2011) Is market fragmentation harming market quality? J Financ Econ 100(3):459–474. https://doi.org/10.1016/j.jfineco.2011.02.006
Sabzian H, Shafia MA, Bonyadi Naeini A, Jandaghi G, Sheikh MJ (2018) A review of agent-based modeling (ABM) concepts and some of its main applications in management science. Iran J Manag Stud 11(4):659–692. https://doi.org/10.22059/ijms.2018.261178.673190
Schelling TC (2006) Micromotives and macrobehavior. W. W. Norton & Company
Scopino G (2016) Do automated trading systems dream of manipulating the price of futures contracts? policing markets for improper trading practices by algorithmic robots. Florida Law Rev 67:221. https://scholarship.law.ufl.edu/flr/vol67/iss1/5
Sewell M (2011) Characterization of financial time series. Research Note, University College London, Department of Computer Science (RN/11/01). https://finance.martinsewell.com/stylized-facts/
Stevens H (2020) Why outbreaks like coronavirus spread exponentially, and how to “flatten the curve”. https://www.washingtonpost.com/graphics/2020/world/corona-simulator/
Takayasu H, Miura H, Hirabayashi T, Hamada K (1992) Statistical properties of deterministic threshold elements — the case of market price. Phys A Stat Mech Appl 184(1):127–134. https://doi.org/10.1016/0378-4371(92)90161-I
Todd A, Beling P, Scherer W, Yang SY (2016) Agent-based financial markets: A review of the methodology and domain. In: 2016 IEEE symposium series on computational intelligence (SSCI), pp 1–5. https://doi.org/10.1109/SSCI.2016.7850016
Trichet JC (2010) Reflections on the nature of monetary policy non-standard measures and finance theory. https://www.ecb.europa.eu/press/key/date/2010/html/sp101118.en.html
Vyetrenko S, Byrd D, Petosa N, Mahfouz M, Dervovic D, Veloso M, Balch TH (2019) Get real: Realism metrics for robust limit order book market simulations. https://arxiv.org/abs/1912.04941
Weisberg M (2012) Simulation and similarity: Using models to understand the world. Oxford Studies in the Philosophy of Science
Yagi I, Hoshino M, Mizuta T (2020) Analysis of the impact of maker-taker fees on the stock market using agent-based simulation. https://arxiv.org/abs/2010.08992
Yamamoto R (2021) Predictor choice, investor types, and the price impact of trades on the Tokyo stock exchange. Computational Economics. https://doi.org/10.1007/s10614-020-10084-4
Yang X, Zhang J, Ye Q (2020) Tick size and market quality: Simulations based on agent-based artificial stock markets. Intell Syst Account Finance Manag. https://doi.org/10.1002/isaf.1474
Zhao R, Cui Y, Liu X (2020) Tick size and market quality using an agent-based multiple-order-book model. Front Phys 8:135. https://doi.org/10.3389/fphy.2020.00135
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2022 The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
About this chapter
Cite this chapter
Mizuta, T. (2022). Artificial Intelligence (AI) for Financial Markets: A Good AI for Designing Better Financial Markets and a Bad AI for Manipulating Markets. In: Aruka, Y. (eds) Digital Designs for Money, Markets, and Social Dilemmas. Evolutionary Economics and Social Complexity Science, vol 28. Springer, Singapore. https://doi.org/10.1007/978-981-19-0937-5_13
Download citation
DOI: https://doi.org/10.1007/978-981-19-0937-5_13
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-19-0936-8
Online ISBN: 978-981-19-0937-5
eBook Packages: Economics and FinanceEconomics and Finance (R0)