Simulated classical tests in multinomial probit models

This paper compares the application of different versions of the simulated counterparts of the Wald test, the score test, and the likelihood ratio test in one- and multiperiod multinomial probit models. Monte Carlo experiments show that the use of the simple form of the simulated likelihood ratio test delivers relatively robust results regarding the testing of several multinomial probit model specifications. In contrast, the inclusion of the Hessian matrix of the simulated loglikelihood function into the simulated score test and (in the multiperiod multinomial probit model) particularly the inclusion of the quasi-maximum likelihood theory into the simulated likelihood ratio test leads to substantial computational problems. The combined application of the quasi-maximum likelihood theory with the simulated Wald test or the simulated score test is not systematically superior to the application of the other versions of these two simulated classical tests either. Neither an increase in the number of observations nor in the number of random draws in the incorporated Geweke-Hajivassiliou-Keane simulator systematically lead to more precise conformities between the frequencies of type I errors and the basic significance levels. An increase in the number of observations only decreases the frequencies of type II errors, particularly regarding the simulated classical testing of multiperiod multinomial probit model specifications.     

In Good Times but Not in Bad: The Role of Managerial Discretion in Moderating the Stakeholder Management and Financial Performance Relationship

We examine the role of managers in controlling the positive impact of stakeholder management (SM) on firm financial performance (FP) in the long term. We develop and test competing hypotheses on whether managers act as “good citizens” or engage in “self‐dealing” when allowed greater discretion. We test our assertions using dynamic panel data analysis of a sample of 806 U.S. public firms operating in 34 industries over 5 years (2005–2009). Our results indicate a nuanced influence of managerial discretion contexts on the SM‐FP relationship. We infer that given more latitude in decision making, as long as the “going is good” managers act as good citizens, but otherwise they revert to managerial self‐dealing. In light of our results, firms designing governance mechanisms to encourage managers to balance the needs of both shareholders and stakeholders must remain cognizant of contextual contingencies.     

Unemployment Insurance in Algeria: Help or Hindrance?

The issue of employment has become a crucial one in Algeria. Not only does the economy no longer offer opportunities for jobseekers; economic reforms have led to countless public-sector businesses being wound up, putting thousands of people out of work. Against this background, a system of temporary unemployment benefits has been introduced, though its effectiveness has been sharply criticized. How many of the occupations targeted by the legislation setting up the system are actually covered by Algeria's unemployment insurance system, and on what terms? What obstacles are there to the full realization of this system's objectives, and how effective is it in an economic and financial sense? The unemployment insurance system is traditionally accused of clogging up the machinery of employment and preventing a return to a competitive labour market. There are also those who believe that this system helps to foster voluntary unemployment and encourages people to spend more time on leisure pursuits. Our article considers all these views. At the end the reader will also find brief details of unemployment insurance in South Africa and Egypt.     

Older people, work and equal opportunity

People in industrialized countries are living longer, having fewer children and retiring sooner. The proportion of older people in the population is rising; their labour force participation is declining. These trends are arousing apprehensions about future labour supply and pension financing as contributor/beneficiary ratios decrease. A widely advocated policy response is to seek higher participation rates, especially by raising pensionable ages and restricting early retirement. Little has been done to improve the labour market position of seniors. Three-pronged action is needed: promoting training, placement and workplace flexibility; changing attitudes through education and information; and prohibiting age discrimination in employment.     

Development Disparities and Peculiarities: An Empirical Investigation

The vast, sprawling and densely populated province of Punjab, while predominantly agricultural, is marked by considerable regional disparities. An attempt has been made in this paper to see to what extent the differences in the human and social development in its various districts is there. The study exhibits that those regions which are ranked as lagging districts have high disparities in terms of human development and social development with high urban and rural poverty estimates. It is clearly revealed that aggregate national figures mask vast sub regional disparities and speedy development in some regions may not necessarily elevate lagging regions.     

Semi-doubly optimal concentric circles fitting with presence of heteroscedasticity

Previous studies focus on homogeneous and isotropic assumptions about the noisy data. Many methods have been developed recently for fitting concentric circles to data. In this paper, these statistical assumptions have been relaxed. To the best of our knowledge, only one iterative method has been recently developed. Due to its complexity, no such algorithm is available to compute the reliable maximum likelihood estimator (MLE). Accordingly, we have developed four new methods that outperform the existing methods including the orthogonal distance regression (ODR). We also discuss which of these methods is superior according to the four principles: statistical efficiency, accuracy, robustness, and computational efficiency. Numerical experiments on synthetic and real images have been conducted to validate our findings.     

A penalized likelihood method for nonseparable space–time generalized additive models

In this paper, we study space–time generalized additive models. We apply the penalyzed likelihood method to fit generalized additive models (GAMs) for nonseparable spatio-temporal correlated data in order to improve the estimation of the response and smooth terms of GAMs. The results show that our space–time generalized additive models estimated response and smooth terms reasonable well, and in addition, the mean squared error, mean absolute deviation and coverage intervals improved considerably compared to the classic GAM. An application on particulate matter concentration in the North-Italian region of Piemonte is also presented.     

FDA: strong consistency of the kNN local linear estimation of the functional conditional density and mode

In this paper we present a new estimator of the conditional density and mode when the co-variables are of functional kind. This estimator is a combination of both, the k-Nearest Neighbours procedure and the functional local linear estimation. Then, for each statistical parameter (conditional density or mode), results concerning the strong consistency and rate of convergence of the estimators are presented. Finally, their performances, for finite sample sizes, are illustrated by using simulated data.     

Normalité asymptotique d'estimateurs convergents du mode conditionnel

We state sufficient conditions for asymptotic normality of convergent estimates of the conditional mode, irrespective of data dependence, and give an application to α-mixing stationary processes.     

Inferences on the ratio of two generalized variances: independent and correlated cases

Statistical inferences about the dispersion of multivariate population are determined by generalized variance. In this article, we consider constructing a confidence interval and testing the hypotheses about the ratio of two independent generalized variances, and the ratio of two dependent generalized variances in two multivariate normal populations. In the case of independence, we first propose a computational approach and then obtain an approximate approach. In the case of dependence, we give an approach using the concepts of generalized confidence interval and generalized p value. In each case, simulation studies are performed for comparing the methods and we find satisfactory results. Practical examples are given for each approach.