1,721,084 research outputs found
Danse avec les marchés : trading algorithmique et émergence rythmique
No full textCette thèse examine les dynamiques complexes de la diffusion de l'information sur les marchés financiers, marquées par une transition majeure du système de cotation à la criée vers des plateformes entièrement électroniques. Guidée par trois questions de recherche interdépendantes, notre étude vise à proposer une compréhension exhaustive de cet écosystème en évolution.Le chapitre 2 propose une exploration empirique pour établir un « modèle rythmique » de diffusion de l'information sur les marchés. Grâce à une analyse rigoureuse des données de trading à haute fréquence, nous démontrons une constance dans la diffusion de l'information, indépendamment de l'actif financier ou de l'année considérés. Ce modèle révèle que l'information se transmet en groupes de cinq événements en moyenne, rappelant des structures rythmiques observées dans la musique et la nature, et suggérant une propriété fondamentale du marché. Cette observation est reliée à des théories cognitives comme la loi de Miller et le problème d'El Farol, soulignant leur pertinence pour comprendre les mécanismes d'interaction algorithmique dans les marchés financiers.Le chapitre 3 se penche sur les bases théoriques de ce « rythme du marché », en questionnant si celui-ci découle d'une efficacité optimale de l'information ou de contraintes opérationnelles. En conceptualisant le marché comme un canal de communication, nous établissons des limites inférieure et supérieure à sa réflexivité, en exploitant des principes de la théorie de l'information et de la thermodynamique stochastique. La limite inférieure est contrainte par les coûts énergétiques liés au traitement rapide de l'information, tandis que la limite supérieure est limitée par le bruit introduit par le trading algorithmique, qui affecte la capacité de transmission de l'information. Nous introduisons également un nouvel indice, l'Energy-Information Efficiency (EIE), pour évaluer les coûts énergétiques du traitement de l'information, et observons que et constatons que les titres « blue chip », souvent associés au trading à haute fréquence, sont les plus efficients selon cette mesure.Dans le chapitre 4, nous explorons l'impact de l'hétérogénéité des plateformes d'échange sur la diffusion de l'information dans les marchés fragmentés. À travers une modélisation par agents, nous démontrons que les caractéristiques variées de ces plateformes, notamment leur réactivité, influencent significativement la réflexivité du marché. Nous constatons que les plateformes très réactives favorisent une réflexivité accrue, contribuant à l'équilibre optimal entre capacité du canal et coût énergétique identifié dans le chapitre précédent. Ce résultat souligne l'importance de prendre en compte les spécificités des plateformes lors de l'évaluation de l'impact du trading algorithmique sur l'efficience du marché.Globalement, cette thèse propose une exploration alternative de la diffusion de l'information et de la formation des prix sur les marchés financiers à haute fréquence. Nos résultats contribuent au corpus croissant de littérature sur la microstructure des marchés à l'ère du trading algorithmique, en offrant de nouvelles perspectives sur l'interaction complexe entre les acteurs du marché, la technologie et la dynamique de l'information. L'émergence d'un rythme de marché constant, ses fondements théoriques et le rôle des infrastructures dans sa formation proposent une nouvelle perspective sur l'évolution et l'impact énergétique des marchés financiers à l'ère numérique.Abstract: This dissertation explores the intricate dynamics of information diffusion in high-frequency financial markets, where the transition from traditional floor trading to fully electronic platforms has dramatically reshaped the landscape. Our investigation is guided by three key research questions, each building upon the previous one to provide a comprehensive understanding of this evolving ecosystem. In Chapter 2, we embark on an empirical journey to uncover the rhythmic patterns of information dissemination in the stock market. Through a rigorous analysis of high-frequency trading data, we reveal a consistent pattern of information diffusion across all assets and several time periods. This pattern, characterized by an average of five events per cluster, resonates with similar rhythmic structures found in music and nature, suggesting a fundamental property of the market. We connect this observation to cognitive theories like Miller’s law and the El Farol bar problem, highlighting the potential for cognitive-like constraints in algorithmic interactions within financial markets. Chapter 3 probes the theoretical underpinnings of this observed market rhythm, exploring whether it arises from optimal information efficiency or operational constraints. By framing the market as an information channel and leveraging concepts from information theory and stochastic thermodynamics, we establish both a lower and upper bound on market reflexivity. The lower bound is dictated by the energetic costs of rapid information processing, while the upper bound is determined by the limits of information transmission capacity due to the noise introduced by algorithmic trading. We further propose a novel metric, Energy-Information Efficiency (EIE), to assess the balance between information processing and energetic costs in different stocks. We find that blue-chip stocks, often associated with high-frequency trading, are the most efficient under this metric. In Chapter 4, we turn our attention to the role of exchange heterogeneity in shaping information diffusion within fragmented markets. After calibrating it from empirical data, we demonstrate with an agent-based model that the diverse characteristics of exchanges, particularly their responsiveness, significantly influence market reflexivity. We find that highly reactive exchanges contribute to greater reflexivity, consistent with the optimal balance between channel capacity and energetic costs identified in Chapter 2. This highlights the importance of considering exchange speed when evaluating the impact of algorithmic trading on market efficiency. Overall, this dissertation offers a comprehensive exploration of information diffusion and price formation in high-frequency financial markets. Our findings contribute to the growing body of literature on market microstructure in the age of algorithmic trading, providing new insights into the complex interplay between market participants, technology, and information dynamics. The emergence of a consistent market rhythm, its theoretical underpinnings, and the role of infrastructure speed in shaping it, offer a novel perspective on the evolution and energetic impact of financial markets in the digital age
Théorie cumulative des perspectives pour le risque et l'incertitude : Nouvelles méthodes de mesure et applications
No full text. La thèse se situe à l'intersection de l'économie comportementale, de l'économie expérimentale et de la théorie de la décision. Les chapitres 1, 2 et 3 développent des méthodes pour estimer les différentes composantes des modèles de décision en situation de risque et d'incertitude : fonction d'utilité, fonction de pondération, aversion aux pertes et croyances. Les applications confirment des déviations par rapport aux théories standard (utilité espérée et utilité espérée subjective) à travers la fonction de pondération, l'aversion aux pertes et les attitudes d'ambiguïté. Les gens sont plus insensibles à la probabilité en présence d'événements asymétriques qu'en présence d'événements symétriques, ce qui suggère que la formation des croyances demande des efforts cognitifs. Pour une même source d'incertitude, les individus font preuve d'aversion à la dépendance des gains et de préférence pour la variété des gains. L'aversion à la dépendance des gains signifie que les individus n'aiment pas que leurs gains dépendent des préférences des autres. Ce comportement se traduit par une fonction d'utilité concave. La préférence pour la variété des gains signifie que les individus préfèrent un plus grand nombre de possibilité de gains lorsque les gains dépendent des préférences des autres. Ce comportement se traduit par l'optimisme. Le chapitre 4 étudie l'existence de l'arbitrage entre risque et incitations (RIT) dans le cadre de l'utilité dépendante du rang (RDU) et de la moyenne-variance-skewness (MVS). Les analyses théoriques montrent que le RIT est remarquablement robuste sous RDU mais pas sous MVS. Avec des données basées sur un nouveau modèle expérimental qui élimine les facteurs de confusion, le chapitre 4 fournit des preuves de l'existence du RIT même dans le cas où les individus ont des préférences pour le risque, ce qui est une prédiction distincte du RDU. Les résultats confirment l'existence du RIT et suggèrent qu'il s'applique à un large éventail de situations, y compris les cas où les individus ont des préférences pour le risque (par exemple, la rémunération des dirigeants).The thesis is at the intersection of behavioral economics, experimental economics, and decision theory. Chapters 1, 2, and 3 develop methods to estimate different components of decision models under risk and uncertainty: utility function, weighting function, loss aversion, and beliefs. Applications confirm deviations from standard theories (Expected Utility and Subjective Expected Utility) through evidence of weighting function, loss aversion, and ambiguity attitudes. People are more insensitive to likelihood in the presence of asymmetric events than symmetric events, suggesting that belief formation is cognitively demanding. For equal sources of uncertainty, people exhibit payoff dependence aversion and variety of payoffs seeking. Payoff dependence aversion means that people dislike that their own payoffs depend on the preferences of others. This behavior is captured by a more concave utility function. Variety of payoffs seeking means that subjects prefer a greater number of possible payoffs when such possible payoffs depend on the preferences of others. This behavior is captured by more optimism. Chapter 4 studies the existence of the Risk-incentives tradeoff (RIT) under Rank Dependent Utility (RDU) and Mean-Variance-Skewness (MVS). Theoretical analyses show that RIT is remarkably robust under RDU but not under MVS. With data based on a novel experimental design that eliminates confounding factors, chapter 4 provides evidence for RIT even in the case of risk-seeking agents, which is a distinct prediction of RDU. The results provide support for the RIT and suggest that it applies to a broad range of situations, including cases in which agents are risk-seeking (e.g., executive compensation)
Watching or not watching? Access to information and the incentive effects of firing threats
Full text linkA common rationale for the use of salary contracts is that they can produce substantial incentive effects when coupled with firing threats. However, enforcing firing threats may require close supervision of employees, thus possibly offsetting the very reasons salaries are commonly used, such as lowering monitoring costs and granting autonomy to employees. We design a series of experiments to study the effectiveness of firing threats when only limited information is available to supervisors. We show that light and unobtrusive supervision can produce large incentive effects. Compared to salary contracts, firing threats based on observing organizational performance alone increase employees’ output by 70% whereas only observing how long an employee works doubles output. These findings show that salaries can produce large incentive effects even in the absence of intensive supervision. Finally, we show that salary contracts with firing threats perform at least as well as other popular incentive schemes, such as bonuses, individual and team incentives, that rely on a similar amount of information about employees.The authors gratefully acknowledge financial support from the Spanish Ministry of Economics and Competitiveness through Grant: ECO2017–88130 and through the Severo Ochoa Program for Centers of Excellence in R&D (CEX2019–000915-S), the Generalitat de Catalunya (Grant: 2017 SGR 1136).Peer reviewe
An Experimental Test of Algorithmic Dismissals
Full text linkWe design a laboratory experiment in which a human or an algorithm decides which of two workers to dismiss. The algorithm automatically dismisses the least productive worker whereas human bosses have full discretion over their decisions. Using performance metrics and questionnaires, we find that fired workers react more negatively to human than to algorithmic decisions in a broad range of tasks. We show that spitefulness exacerbated this negative reaction. Our findings suggest algorithms could help tame negative reactions to dismissals
Peer Evaluations And Team Performance: When Friends Do Worse Than Strangers
No full textWe use peer assessments as a tool to allocate joint profits in a real-effort team experiment. We find that using this incentive mechanism reduces team performance. More specifically, we show that teams composed of acquaintances rather than strangers actually underperform in a context of peer evaluations. We conjecture that peer evaluations undermine the inherently high level of intrinsic motivation that characterizes teams composed of friends and possibly exacerbate negative reciprocity among partners. Finally, we analyze the determinants of peer assessments and stress the crucial importance of equality concerns
Going Beyond Counting First Authors in Author Co-citation Analysis
Full text linkThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
Team formation and biased self-attribution
Full text linkWe analyze the impact of individuals' self-attribution biases on the formation of teams in the workplace. We consider a two periods model in which workers jointly decide whether to form a team or work alone. We assume workers' abilities are unknown. Agents update their beliefs about abilities after receiving a signal at the end of the first period. We show that allowing workers to learn about their abilities undermines cooperation when a fixed allocation of the group outcome is assumed. Consistent with the latter finding, we establish that making learning about workers' abilities less accessible increases workers' cooperation and welfare. When workers suffer from selfserving attribution, cooperation among agents is undermined whatever the allocation rule considered for the group outcome. We analyze possible solutions to insufficient teamwork. We find that team contracts based on a revelation game can improve cooperation as well as the presence of a manager in the team. Full efficiency is however never achieved. Our paper establishes a basic framework to analyze necessary psychological conditions for individuals to form teams. We apply our model to coauthorship and to organizational issues
- …
