Robust m-estimators

This paper provides a summary of the influence function approach to robust estimation of parametric models. Hampel's optimality results for M-estimators with a bounded influence function is generalized to allow for arbitrary choices of the asymptotic efficiency criterion and the norm of the influence function. Further extensions to other cases of practical interest are also considered.     

Technology for social housing in Latin America : An evaluation of the CYTED research and development programme

Since the mid 1980s an international programme of scientific and technological co-operation between Spain, Portugal and Latin America has been underway (CYTED). One of the 16 sub-programmes is focussing on Technology for Social Housing, and has brought together hundreds of researchers and housing activists from every corner of the continent, to seek pragmatic approaches to improve the housing conditions of the poor majority in Latin America. In addition to cataloguing and exchanging existing experiences, a wide-ranging and detailed programme of research and development of innovative housing technologies has been set in motion. These include materials, techniques and processes.The paper briefly outlines the characteristics of the existing housing situation in Latin America in order to identify the main elements upon which these new approaches are built. The key components of the CYTED housing programme are then described and illustrated using examples of different projects. A common theme is the relationship between the building process and social patterns, particularly the potential for cooperation and community self-help construction. The paper then identifies some limitations of the programme and concludes that despite the commitment and creativity of those involved, the effectiveness of many projects and the general validity of the approaches adopted, severe structural constraints are likely to limit the impact of the programme. However the increased contacts, communication and opportunities for collaboration within the continent provide grounds for limited optimism about future developments.     

Gambling Screens: Does Shortening the Time Frame Affect their Psychometric Properties?

Gambling screening tools such as the South Oaks Gambling Screen (SOGS) and a DSM-IV Screen for Gambling Problems (NODS) developed by the National Opinion Research Council have psychometrically validated lifetime and past-year versions. As research questions often dictate shorter time intervals, researchers have adapted the time frames of these instruments to their specific purposes without examining whether changing the time frame affects the psychometric properties. In this study, 3-month versions of SOGS and NODS were administered to a sample of 80 pathological gamblers (59 men, 21 women, mean age 44) seeking treatment at a state-funded facility. The 3-months versions had good internal consistency, good convergent validity with each other, with gambling behaviors assessed via the timeline followback method, and with measures of impulsivity. The 3-month versions also showed good discriminant validity with demographic variables and a measure of verbal IQ. Together the data indicate that shortening the time frame to 3 months does not seem to have adverse effects on the psychometric properties of SOGS and NODS. Thus these adapted versions could profitably be used for shorter time intervals, including as pre/post-treatment and follow-up measures in treatment outcome studies.     

A Note on the Portfolio Selection Problem

In this note we provide new results of interest in the portfolio choice problem when the risky opportunities are correlated: for a general vector (X₁, X₂,..., X_n) of risky opportunities we give new conditions for stochastic comparison among different portfolios choices and new necessary and sufficient conditions to characterize the portfolio which gives the maximal expected utility.     

Total factor productivity and shareholder returns in banking

This paper examines shareholder value drivers in European banking focusing on the efficiency and productivity features of individual banks. In particular, we analyse the value relevance of bank cost efficiency and total factor productivity (TFP) (in all its components, including technological change, pure technical efficiency change and scale efficiency change) to see how these influence shareholder value creation in European banking. The paper focuses on the French, German, Italian and U.K. banking systems over the period 1995–2002 and includes both listed and non-listed banks. We find that TFP changes best explain variations in shareholder value (measured by market-adjusted returns, MAR, for listed banks and by the ratio of EVA_bkg to invested capital at time t−1 for non-listed banks). In both samples, we also find that technological change seems to be the most important component of TFP influencing shareholder value creation in European banking.     

Generalized Mixtures of Gamma and Exponentials and Reliability Properties of the Maximum from Friday and Patil Bivariate Exponential Model

The minimum and maximum order statistics from many of the common bivariate exponential distributions are predominantly generalized mixtures of exponentials; however, the maximum from the Friday and Patil bivariate exponential (FPBVE) model is either a generalized mixture of three or fewer exponentials or a generalized mixture of gamma and exponentials. In this article, we obtain conditions based on the weights and parameters of the generalized mixtures of gamma and one or two exponential distributions that yield legitimate probability models. Furthermore, we analyze properties of the failure rate of the maximum from the FPBVE model. This answers a question raised in Baggs and Nagaraja (1996 Baggs , G. E. , Nagaraja , H. N. ( 1996 ). Reliability properties of order statistics from bivariate exponential distributions . Commun. Statist. Stochastic Mod. 12 : 611 – 631 .[Taylor & Francis Online] , [Google Scholar]).     

Stochastic Aging Classes for the Maximum Statistic from Friday and Patil Bivariate Exponential Distribution Family

Friday and Patil bivariate exponential (FPBVE) distribution family is one of the most flexible bivariate exponential distributions in the literature; among others, it contains the bivariate exponential models due to Freund, Marshall–Olkin, Block–Basu, and Proschan–Sullo as particular cases. In this article, we discuss the stochastic aging of the maximum statistic from FPBVE model in according to the log-concavity of its density function, i.e., in the increasing or decreasing likelihood ratio classes (ILR or DLR), and consequently in the IFR and DFR classes. Furthermore, a kind of DFR distributions which are not DLR is derived from our classification.     

Confidence Intervals for the Hyperparameters in Structural Models

This article deals with the bootstrap as an alternative method to construct confidence intervals for the hyperparameters of structural models. The bootstrap procedure considered is the classical nonparametric bootstrap in the residuals of the fitted model using a well-known approach. The performance of this procedure is empirically obtained through Monte Carlo simulations implemented in Ox. Asymptotic and percentile bootstrap confidence intervals for the hyperparameters are built and compared by means of the coverage percentages. The results are similar but the bootstrap procedure is better for small sample sizes. The methods are applied to a real time series and confidence intervals are built for the hyperparameters.     

Book reviews

The World of the Passions in the Utopian Imaginary (Symposium on Utopia) Bruna Consarelli and Nicola Di Penta (eds) Giuffrè Editore, Milano, 1997, pp. 405

Giambattista Vico in the Naples and Seville Centres