PurposeThe purpose of this paper is to identify efficient governance using a governance efficiency score based on recommendations provided by codes of best practices in order to determine “good governance”.Design/methodology/approachBased on a sample of 320 US listed firms from 1994‐2001, governance practices were synthesized by an index computed according to a parametric method, the stochastic frontier analysis, which allows taking into account the relation between inputs (governance axes) and outputs (performance).FindingsThe use of a latent classis in the specification of the model allowed detecting two groups of firms according to their specific characteristics. The results of affectation equation show that the probability of being in the highest performing group is more important when the firm size, the dividend yield and the return on equity (ROE) are high, while a high leverage level decreases the chance to be in the non‐performing group. Moreover, the model allows establishing a dualist description of the two groups which point out two opposite governance systems. The non‐performing system is characterized by a managerial discretion, an ownership concentration, a dominance of the board by the CEO and a manager entrenchment. However, the highest performing system is characterized by an inside control efficiency and an inside financial control efficiency.Research limitations/implicationsThe sample choice presents a selectivity bias. Firms of the sample present some particularities in relation to other US firms, which limits the study generalisation. This study can also be the object of replications in other contexts.Originality/valueThis work is a demarcation in relation to previous works studying corporate governance quality, and particularly the relation between governance and performance. It provides a new econometric approach to develop a synthetic index to evaluate corporate governance firms' practices, wedged on performance level achieved by different firms.

No abstract

Les modèles de marché du travail dans les économies en développement ont généralement privilégié une interprétation en termes de dualisme salarial. L'hypothèse a déjà été élaborée dans les modèles de première génération (Lewis, 1954) et développée à la suite des travaux de Harris-Todaro (1970). L'analyse économétrique est venue par la suite confirmer l'existence de différentiels de salaires entre catégories de travail a priori équivalent. Le test de l'hypothèse de dualisme salarial en Tunisie est effectué dans le cadre de traitements de données visant à lever le biais de sélection. Il s'agit d'aller au-delà des tests courants et déjà existants, qui consistent généralement à régresser séparément et au moyen des MCO, une fonction de gains dans les deux secteurs déterminés a priori. Un modèle à deux régimes permet alors une estimation non biaisée. Par ailleurs, il reste encore à lever l'arbitraire du choix du critère de sélection. Le modèle à régression discriminante endogène non observable s'avère particulièrement adapté à l'opération. On vérifie ici notamment l'existence d'une faible mobilité entre les différents secteurs d'emplois.

No abstract

This chapter examines the antecedents and consequences of the perceived risk of investors towards the Tunisian stock market. A questionnaire was developed and distributed to 411 individual investors chosen by 24 brokerage firms. Using the structural equation model, we operationalize the risk following the psychometric paradigm according to subjective variables (i.e. familiarity and controllability). Results prove that controllability is a significant factor in the formation of perceived risk. We also show that several factors related to the investor, the listed companies and to the stock market can influence the perceived risk by the investor towards the Tunisian stock market. Similarly, we find that perceived risk leads to intensive information search, good performance and a strong reinvestment intention. These results attest the importance of the risk perception in the decision-making process.