Olivier Brandouy

A Generic Architecture for Realistic Simulations of Complex Financial Dynamics

Artificial Stock Markets (here after ASM) have received an increasing amount of academic interest since the seminal works of [20] or [17]. Such platforms have benefitted from advances and new methods developed in the field of multi agentsystems (see for example [10], [14] and [27]). These Agent-Based virtual environments are particularly useful to study various aspects of the financial world in an entirely controlled environment, opening new perspectives for policy makers, regulatory institutions and firms developing business solutions in the financial industry (for example asset management or trading). There is little doubt ASM could become a key system in the post financial crisis risk-management toolbox to overcome the weaknesses of traditional approaches. Agents-based modeling and simulation offer frameworks to study the impact of a Tobin’s tax for example, or to develop new stress tests for assessing financial resilience to economic shocks or to develop new automatic trading techniques.

Testing Double Auction as a Component Within a Generic Market Model Architecture

Since the first multi-agents based market simulations in the nineties, many different artificial stock market models have been developped. There are mainly used to reproduce and understand real markets statistical properties such as fat tails, volatility clustering and positive auto-correlation of absolute returns. Though they share common goals, these market models are most of the time different one from another: some are based on equations, others on complex microstructures, some are synchronous, others are asynchronous. It is hence hard to understand which characteristic of the market model used is at the origin of observed statistical properties. To investigate this question, we propose a generic model of artificial markets architecture which allows to freely compose modules coming from existing market models. To illustrate this formalism, we implement these components to propose a model of an asynchronous double auction based on an order-book and show that many stylized facts of real stock markets are reproduced with our model.

Frontier-based vs. traditional mutual fund ratings: A first backtesting analysis

We explore the potential benefits of a series of existing and new non-parametric convex and non-convex frontier-based fund rating models to summarize the information contained in the moments of the mutual fund price series. Limiting ourselves to the traditional mean-variance portfolio setting, we test in a simple backtesting setup whether these efficiency measures fare any better than more traditional financial performance measures in selecting promising investment opportunities. The evidence points to a remarkable superior performance of these frontier models compared to most, but not all traditional financial performance measures.

Efficient Monitoring of Financial Orders with Agent-Based Technologies

The execution of orders on stock exchanges is managed by a set of formalized rules based on price and time priority. Nevertheless, orders issued by investors do not show-up directly in the market system : they transit through the brokerage intermediation where they can be arranged in different sequences. We show that the latter operation has a critical impact on investors. In this paper, we propose a decision support system that solves the underlying optimization problem for a given social welfare. We show that the solution cannot be obtained without an agent-based simulation platform that individualizes the consequences of the broker decision in terms of order sequencing at the agent (client) level. In this framework, we study the impact of several social welfares functions and show how the broker can grant his clients with ”just and equitable principles of trade”.

On the Design of Agent-Based Artificial Stock Markets

The purpose of this paper is to define software engineering abstractions that provide a generic framework for stock market simulations. We demonstrate a series of key points and principles that has governed the development of an Agent-Based financial market application programming interface (API). The simulator architecture is presented. During artificial market construction we have faced the whole variety of agent-based modelling issues : local interaction, distributed knowledge and resources, heterogeneous environments, agents autonomy, artificial intelligence, speech acts, discrete or continuous scheduling and simulation. Our study demonstrates that the choices made for agent-based modelling in this context deeply impact the resulting market dynamics and proposes a series of advances regarding the main limits the existing platforms actually meet.

A Conceptual Framework for the Evaluation of Agent-Based Trading and Technical Analysis

The major part of research dedicated to technical analysis and active trading (i.e., the management of financial portfolios using chartism or moving average indicators for instance) generally focuses on single “signals” giving the opportunity to buy or sell a financial commodity frequently a well diversified portfolio (see the extensive survey of |Park and Irwin, 2004). In this context, it has been extensively argued that technical analysis is useless in order to outperform the market (Jensen and Benington, 1969). The reason for that is, assuming informational efficiency (Fama, 1970), all relevant piece of information is instantaneously aggregated in prices. Therefore, there is nothing to extract from previous quotations relevant for one willing to trade on this basis. Since information is, by definition, unpredictable, next price fluctuations will be driven by innovation and the price motion will fluctuate randomly as a result. Nevertheless, empirical investigations tackling this question of “technical trading” exhibit heterogeneous results. On the one hand, a large part of these researches shows that, once risk taken into account, no-one can seriously expect any rate of return over what can be earned with a simple Buy and Hold strategy (henceforth BH |Dempster and Jone, 2005; Detry and Gregoire, 2001).

Optimal Portfolio Diversification? A multi-agents ecological competition analysis

In this research we study the relative performance of investment strategies scrutinizing their behaviour in an ecological competition where populations of artificial investors co-evolve.We test different variations around the canonical modern portfolio theory of Markowitz, strategies based on the naive diversification principles and the combination of several strategies.We show, among others, that the best possible strategy over the long run always relies on a mix ofMean-Variance sophisticated optimization and a naive diversification. We show that this result is robust when short selling is allowed in the market and whatever the performance indicator chosen to gauge the relative interest of the studied investment strategies.

L'apport des SMA à la modélisation des marchés financiers

We study in this article several multi-agents models of stock markets. We present in a first part models classicly used by economists to simulate markets, which use a central equation to aggregate agents decisions, which implies that the market is synchronous and non continuous. We show that these kind of models are not sufficient in order to reproduce financial markets behaviors, and present our own orderbook-based market model. We show that reproducing the interaction model used by agents on real markets allows to obtain better results, and allows to better understand how statistical properties observed on real markets emerge from agents interactions through the market.

Asymmetric information, imitative behaviour and communication: price formation in an experimental asset market

The paper describes experimental results from a simulated stock market with manipulation of asymmetric information and communication, including conditions intended to promote imitative behaviour and rumour. Price discovery was inefficient when the presence of insiders was disguised, compared to a homogeneous expectations baseline. When the presence of insiders was revealed observed prices became efficient with respect to bad news but not with respect to good news, possibly suggesting loss-averse behaviour. With free communication there was a decrease in both efficiency and price volatility—insider information was masked by noise. Price formation under these conditions was similar to baseline, but with weak evidence of speculative pricing. It is conjectured that other factors than informational noise may be necessary determinants of herd behaviour, but further work is indicated.

A computational definition of financial randomness

Under the Efficient Market Hypothesis (EMH), in a one security and cash trading system, no one can outperform the buy-and-hold strategy in the long run. This classical proposition calls for a formal definition of outperforming trading rules. This article proposes a computational approach that completely departs from the probabilistic framework in which the question was originally formulated. Inspired by Schnorr’s definition of binary random string (Schnorr [Math. Syst. Theory, 1971, 5, 246–258]), computable functions are used to model effective trading rules that can be applied to financial price series. In the absence of transaction costs, a price sequence is said to be ‘unbeatable’, if no effective trading rule can generate indefinitely more profits than its buy-and-hold alternative. As a quantitative formulation of recent thoughts on EMH (e.g. Lo [J. Portfolio Manage., 2004, 30, 15–20], Kandhani and Lo [J. Invest. Manage., 2007, 5, 29–78], Hasanhodzic et al. [Quant. Financ., 2011, 11(7), 1043–1050]), our definition reconsiders the notion of financial randomness and reconciles EMH with the performance of ‘constantly re-balanced portfolios’ and ‘automated trading systems’ (Cover [Math. Financ., 1991, 1(1), 1–29], Cover and Ordentlich [IEEE Trans. Inf. Theory, 1996, 42, 348–363], Fagiuoli et al. [Quant. Finance, 2007, 7(2), 161–173], Creamer and Freund [Quant. Finance, 2010, 10(4), 401–420]). Thus, we propose a ‘computational definition of financial randomness’ in formulating the concept of ‘unbeatable price sequences’.