L'Hypothése de la Distribution Normale dans le cas d‘Événements Simultanes

L'hypothèse de la distribution normale lors de l'analyse des données, permet entre autres, l'utilisation des tests statistiques classiques, ce que n'autorise pas une distribution asymétrique ou aplatie au centre et allongée awe extrémités. Ce travails'adresseprécisé-ment à la forme de la distribution des résidus quotidiens dans le cas d'évènements simultanés, e'est-d̀-dire ceux affectant un certain nombre d'entreprises au mǐme moment, puisque aucune étude ne s'est préoccuppée, à notre connaissance, de ce phénomène. Nos résultats semblent indiquer que les hypothèses de la distribution normale et de non-aplatissement des résidus quotidiens dans le cas d'événements simultanés sont irréalistes, le sort de l'hypothèse d'asymétrie ne semblepas clair surtout à un intervalle de condiance de 99 pourcent. L'utilité des tests statistiques classiques se trouve, par conséquent, très minimisée. Ces tests auront, en effet, tendance à rejeter l'hypothèse de performance normale mǐme en l'absence de toute performance anormale. Abstract The hypothesis of normal distribution in the analysis of data permits, among other things, the use of classical statistical tests. This is not allowed by a distribution which is asymetric or flat at the center and extended at the extremities. This paper deals specifically with the shape of the distribution of daily remainders in the case of simultaneous events which affect a number of businesses at the same time because, to our knowledge, no study has previously dealt with this phenomenon. Our results seem to indicate that the hypothesis of normal distribution and of the non-flattening of daily remainders in the case of simultaneous events are unrealistic and the outcome of hypothesis of asymetry does not seem clear especially at a 99% confidence interval. As a result, the use of classical statistical tests is minimized. In effect, these tests will generally reject the hypothesis of normal performance even in the absence of any abnormal performance.     

Effets des réglementations sur la rentabilité et le risque des compagnies d'assurances iard

La réglementation des sociétés d'assurances par les autorités de tutelle impose différentes contraintes aux gestionnaires. Il est difficile d'en apprécier les effets finaux si l'on ne dispose pas d'un modèle d'ensemble de l'entreprise. La rentabilité des compagnies d'assurances dépend de la gestion de l'actif et du passif du bilan, portefeuille déplacements et portefeuille d'assurances. La réglementation intervient pour poser des limites à ces choix sous forme de plafonds en pourcentages pour certains types de placements. Cette étude dans le contexte Canadien met en évidence le coǔt économique, en termes de risque et de rendement, d'une réglementation sur les placements. Abstract The control of insurance companies by the regulatory authorities imposes constraints on the manager's decisions. The companies' profitability is a function of the management of both assets and liabilities. Therefore it depends on the relative importance of the diverse lines of insurance - the insurance portfolio - and relative weight of the various assets - the investment portfolio. Portfolio theory demonstrates that the total risk is a non-linear combination of those component risks in a portfolio. This, in turn, determines the pursuit of the optimal portfolio structure, a structure defined by a minimal risk for a given level of profitability. With such a generalized portfolio model for insurance companies, one could simulate the consequences of the modification of any given regulation on either the assets or liabilities. The combination of risk level and profitability accessible to the companies vary as a function of markets constraints and regulations considered. It is therefore possible to demonstrate the effect, in terms of risk and return framework, of various regulations. The first section of this paper discusses the appropriate approach to the portfolio problem applied to insurance. The second section describes the data and the empirical results.     

Effets des conceptions de l'erreur,de la motivation à apprendre et du soutien social sur l'intention de transférer les nouveaux apprentissages : un modèle de médiation modérée

This study aims to assess whether trainees' errors orientation can influence their motivation to learn and their intention to transfer learning. Data were collected from 275 trainees participating in a corporate training program offered by a Canadian government agency. Results show that the concept of learning by error positively influences trainees' levels of motivation and their intention to transfer learning. Results also confirm the mediating role motivation plays between the concept of learning by error and the intention of transferring learning. Finally, results confirm the moderating role of social support on the relationship between motivation to learn and the intention to transfer learning. Practical and theoretical implications of these results are discussed. Copyright © 2017 ASAC. Published by John Wiley & Sons, Ltd.