Cognition & Economics  (CE) Session 1

Schedule Top Page

Time and Date: 10:45 - 12:45 on 22nd Sep 2016

Room: J - Derkinderen kamer

Chair: Jelena Grujic

130 Will the market go up or down? Human guesses facing market uncertainty in a lab-in-the-field experiment as a way to unveil unintended strategies and behavioral biases [abstract]
Abstract: Decisions taken in our everyday lives are based on a wide variety of information so it is generally very difficult to assess what are the strategies that guide us. Stock markets therefore provides a rich environment to study how people take decisions since responding to market uncertainty needs constant updates of such strategies. For this purpose, we run a lab-in-the-field experiment where volunteers are given a controlled set of financial information -based on real data from worldwide financial indices- and they are required to guess whether the market price would go up or down in each situation. From the data collected we explore basic statistical traits, behavioral biases and emerging strategies. In particular, we detect unintended patterns of behavior through consistent actions which can be interpreted as Market Imitation and Win-Stay Lose-Shift emerging strategies, being Market Imitation the most dominant one. We also observe that these strategies are affected by external factors: the expert advice, the lack of information or an information overload reinforce the use of these intuitive strategies, while the probability to follow them significantly decreases when subjects spends more time to take a decision. The cohort analysis shows that women and children are more prone to use such strategies although their performance is not undermined. Our results are of interest for better handling clients expectations of trading companies, avoiding behavioral anomalies in financial analysts decisions and improving not only the design of markets but also the trading digital interfaces where information is set down. Strategies and behavioral biases observed can also be translated into new agent based modeling or stochastic price dynamics to better understand financial bubbles or the effects of asymmetric risk perception to price drops. Full paper available: Website:
Josep Perello, Jordi Duch and Mario GutiƩrrez-Roig
207 Drunk Game Theory. An individual perception-based framework for evolutionary game theory. [abstract]
Abstract: We present Drunk Game Theory (DGT), a framework for individual perception-based games, where payoffs change according to player's previous experience. We introduce DGT with the narrative of two individuals in a pub choosing independently and simultaneously between two possible actions: offering (cooperating) or not (defecting) a round of drinks. The payoffs of these interactions are perceived by individuals as a function of their current states. We represent these perceptions through two different games. The first game constitutes the classic Prisoner's Dilemma, in which player's utility is a function of the number of consumed drinks and invested money. The second game takes the form of the Harmony game, in which payoffs are computed solely as the number of consumed drinks. Players perceive one of the two games according to their current cognitive level. An individual is more likely to perceive payoffs according to the Prisoner's Dilemma (Harmony) game when she is in a heightened (diminished) cognitive state. The cognitive level of a player evolves according to the outcome of her previous interactions: it reduces when drinking and it increases when abstaining. We use evolutionary game theory to model the evolution of cooperation within well-mixed and structured populations. We observe non-trivial dynamics in both the fraction of cooperators and the cognitive levels when cooperators and defectors dynamically coexist over time. Our analytical results in well-mixed populations agree with those obtained from agent-based simulations. We further explore the role of network-constrained interactions on the overall level of cooperation. By accounting for heterogeneous and feedback-dependent perceptions, the DGT framework opens new horizons for exploring the emergence of cooperation in social environments.
Nicholas Mathis, Leto Peel, Massimo Stella, Luis A. Martinez-Vaquero and Alberto Antonioni
105 Predicting missing links in criminal networks: the Oversize case [abstract]
Abstract: The problem of link prediction in networks has recently received increasing attention. One of its aims is to recover missing links, namely connections among actors which are likely to exist but have not been reported because data are incomplete or uncertain. In criminal investigations, problems of incomplete information are encountered almost by definition, given the obvious anti-detection strategies set up by criminals and the limited investigative resources. We analyze a specific dataset obtained from a real investigation (operation Oversize, an Italian criminal case against a mafia group) including all wiretap conversations recorded by investigators and involving 182 individuals. The peculiarity of the case lies in the availability of three networks derived at different stages of the proceedings: the wiretap records (WR), including all wiretap conversations; the arrest warrant (AW), with a selection of the transcripts; and the judgment (JU), summarizing the trial. A few links are removed passing from WR to AW and to JU, since they are considered not relevant by the investigators (marginal links). We are aimed at identifying missing links, namely highly probable social ties not revealed by wiretap records. We propose a strategy based on the topological analysis of the links classified as marginal, i.e. removed during the investigation procedure. The main assumption is that missing (i.e. undetected) links should have opposite features with respect to marginal ones. Surprisingly, a centrality measure such as link betweenness proves to be unable to characterize marginal links, which are instead captured by standard measures of node similarity. A pool of link prediction methods, both local and global, is then applied and their results are analyzed and compared. The thorough inspection of the judicial source documents validates the results and confirms that the predicted links, in most instances, do relate actors with large likelihood of co-participation in illicit activities.
Giulia Berlusconi, Francesco Calderoni, Nicola Parolini, Marco Verani and Carlo Piccardi
6 The Risky Business of Asking for Help: An Agent Based Model of Unmet Need [abstract]
Abstract: In this work we present an agent based model of elderly care were populations of decision theoretic agents play a game, reflecting the interwoven supply and demand side decision making processes that govern whether older adults seek, and receive support in their activities of daily living. The model draws together longitudinal survey (ELSA) data to provide base rates of need for support, care costs from local authority activity reports (HSCIC PSSE/PSSA), and attitude surveys (ONS OPN, EuroBarometer, and ESS) to produce distributions of synthetic agent psychologies. We then calibrate the model against reported rates of unmet need from the ELSA dataset, by building statistical emulators of the simulation model to rapidly explore the free parameter space. The simulation results suggest that the care system is most sensitive to the balance between the perceived costs of failing to provide care where needed, and the rewards of delivering appropriate support. Further to this, the model indicates that the real system lies near to collapse, with relatively small decreases in perceived costs and rewards leading to breakdown. Potential applications for the simulation itself are in the arena of policy development, by suggesting possible implications for interventions, for example the impact of increases in the cost of care provision, or of campaigns targeting the the perception of stigma attached to age. In addition, the parameterisation and calibration of the model demonstrate the possibilities of simulation as a method for integrating disparate data sources.
Jonathan Gray, Jakub Bijak and Seth Bullock
469 Opinion evolution in the presence of stubborn agents: from consensus to disagreement [abstract]
Abstract: The structural properties of social networks, naturally described by graphs of interpersonal ties, has been thoroughly studied by the interdisciplinary theory of Social Network Analysis (SNA). However, the dynamics and evolution of social systems have mainly remained beyond the scope of SNA, confined to special processes over social networks, such as e.g. random walks and epidemic spread. In spite of the rapid progress in study of complex systems and their dynamics, reinforced by the development of software tools for big data analysis, the realm of dynamic social networks still remains a challenge for the modern science. An important reason for that is the lack of mathematical models, representing social groups by dynamical systems. Such models should be sufficiently simple to allow their rigorous analysis and still remain sufficiently "rich" to capture the behavior of real social systems. Unlike many natural and engineered systems, social networks rarely exhibit regular behaviors like consensus and synchronization; the opinions and actions of social agents are usually featured by persistent disagreement and clustering. In this paper, we consider a model of opinion evolution is social group, proposed by Friedkin and Johnsen in 1999 and confirmed by experiments with small social groups. This model extends the consensus-based procedure for decision making, dating back to the works by French (1956) and DeGroot (1974), to the case where some agents are "stubborn" and "attached" to their initial opinions, factoring them into every stage of the opinion iteration. We offer necessary and sufficient graph-theoretic conditions for the convergence of opinions in the Friedkin-Johnsen model. This model is also extended to the case where opinions are multidimensional and consist of interdependent scalar topic-specific opinions; such model can be used to model the evolution of belief systems.
Anton Proskurnikov, Sergei Parsegov, Roberto Tempo and Noah Friedkin