Contenu du sommaire : Économie de l'éducation

Education and Economic Development, by Edmond Malinvaud. Although education systems frustrate overly high hopes, they are a continual focus of attention. Improving their hard-to-measure efficiency is a particular challenge in that there are many problems to be addressed and the results of educational sciences cannot be easily interpreted in terms of general guidelines. Thirty years ago, economists had a simple, albeit imperfect, methodology for studying the link between education and growth. Interest in new growth theories has led to the reconsideration and profitable extension of former analyses. Yet the main questions that prevailed before unfortunately remain unanswered.

Human Capital and Economic Growth, by Kinvi D.A. Logossah. This paper sets out to review the theory of human capital while examining its empirical validity. It shows that the traditional hypothesis whereby education is an investment that increases individual efficiency and thus productivity and gains has been enlarged by various criticisms such as filter models, theories of labour market heterogeneousness, the efficient wages model, and domination, discrimination and patriarchate hypotheses. Therefore, although the level of education might appear to be a reliable measurement of the qualification level (orthodox model), it could also be regarded as an indicator of potential productivity (filter model). Moreover, other factors besides education influence productivity and individual gains. These are mainly seniority and experience and, more generally, the characteristics of the labour demand (theories of labour market heterogeneousness, the efficient wages model, and domination, discrimination and patriarchate theories). The analysis of human capital continues and is taking diverse routes. It has been greatly enhanced by recent studies on endogenous growth.

Educational Choices, General Equilibrium and Economic Growth, by Antoine d'Autume. A proper understanding of the role of human capital in economic growth is best achieved in a framework where microeconomic educational choices are imbedded in a general equilibrium model. We therefore consider an overlapping generations model, written in continuous time to facilitate the description of life-cycle behaviour, and we show that the possibility of endogenous growth arises quite naturally in such a framework. We then use an extended version of Lucas (1988) to analyze education policies.

Growth and Training: The Role of Education Policies, by Eve Caroli. This paper proposes a preliminary approach to the relation between growth and training policies. We use an extended version of the 1988 R. Lucas model, which was one of the first attempts at making human capital endogenous in a growth model. Our hypothesis is that the productivity of average per-capita training time - in terms of human capital growth rates - is not independent of the characteristics of the education system. We posit that it is an increasing, yet saturating, function of the training rate, which here is the exogenous variable in the education policy. Correlatively, the production function expression takes account of the fact that trainers represent so many staff less in the final production sector. The solving of the model shows that the economy's growth rate is an increasing and then decreasing function of the training rate all the way down the balanced growth path. This means that beyond a certain threshold, the increase in the training rate is counterproductive for growth. There thus exists an optimal training policy. This policy varies with the efficiency of the teachers. It is concluded that the lower the quality of the training staff and/or the poorer the teaching conditions, the lower the maximum growth rate able to be attained by the economy.

The Relation Between the Quality of the Labour Factor and its Productive Efficiency, by Jean-Pierre Huiban. This article studies the link between the quality of the labour factor and its productive efficiency, based on a model in which one production function with a heterogeneous labour factor is estimated for all French industrial sectors. It is then suggested disaggregating the productive system into homogeneous subsets, based on the level of capital intensity followed by the average size of the units, and estimating a specific link for each one of them. The notion of the functional distribution of jobs is thus introduced in order to enhance the definition of the quality of the labour factor. Estimates are made for individual corporate data, first in an intersectorial context and then within the dairy sector. The study concludes that different forms of link exist, within which the relative contribution of the different job categories to the productive efficiency of the whole can vary.

On-the-Job Training and the Return on Experience: A Model of Skills Dissemination, by Louis Lévy-Garboua. This paper points out that Mincerian gains functions do not result from an optimization process and that the interpretation of the facts is based too much on hypotheses (length of finite life, depreciation of human capital and specific training), which diminish the scope. As on-the-job training is the weak link in the theory, we propose a model of on-the-job training through skills dissemination. This model removes the need for the hypotheses. The model limits on-the-job training to the skills contained in the company, which are dictated to the employees by the company rather than chosen freely by the employees. Consequently, instead of emanating from the employee's optimization of his human assets, training is determined by the dissemination of the company's skills. It is therefore shown that the marginal return on experience decreases a lot faster than the quadratic function would suggest and that it is dependent on the relative skills of the company in relation to the new employee. The marginal return on experience is also linked to the rate of skills dissemination. In this model, on-the-job training always decreases with experience and tends towards zero, even though the time scale is infinite, the depreciation in human capital is zero and its returns are constant.

Wage Deviations between the Public and Private Sectors in French-Speaking Africa: A Comparative Analysis, by Jean-Pierre Lachaud. In the structural adjustment programmes implemented in Africa, the idea that public-sector wages are too high compared with private-sector wages forms one of the bases for public-sector wage policies. Yet few studies have actually examined the existence of a sectorial wage differential resulting in a segmentation of the modern sector's labour market in African economies. The author sets out to check this hypothesis using an estimate of gains functions. He bases his work on homogeneous transversal data, taken from recent pilot employment surveys of households in seven French-speaking African capitals, and on longitudinal data inherent to a single country. The main conclusion of this research is that distortions on the modern labour markets of many African countries are not as great as is normally thought. In most of the countries considered, wage deviations between the public and private sectors are essentially due to individual characteristics. This result suggests that the continuation of policies to reduce public-sector wages is no longer totally justifiable, either economically or socially.

Estimating Gains from Panel Data: Endogenous Human Capital and Selection Effects, by Yves Guillotin and Patrick Sevestre. The purpose of this article is to analyze two of the problems encountered when empirically testing gains functions using panel data. First of all, the biases resulting from the correlation between the initial human capital, experience and the unobserved individual characteristics (personal capacities, family-acquired skills, etc.) are contemplated. Secondly, cylinder sampling effects, such as those in the selection process for the individuals in the sample, are also studied. These two problems are illustrated by distinct gains function estimates, obtained using different methods, based on cylinder samples and non-cylinder samples taken from ENSEE's annual wage and salary return panel. The estimates show that appreciable biases can result from not taking the characteristics of the model and the data into account.

An Empirical Evaluation of the Rationality of Belgian Students (1954-1987), by Jean-Luc de Meulemeester. This paper evaluates the extent of economic rationality among young Belgians taking university courses from 1954 to 1987. To do this, a Freeman cobweb model (1989) is estimated using various hypotheses as to the formation of expectations. Arecursive model with adaptive expectations and partial adjustment is selected. The men's expectations are more static than the women's; expectations are not rational. The conclusion considers the irreversible nature of the educational investment and ends on a note of extremely high uncertainty. Consequently, the absence of rational expectations cannot lead to a conclusion of low rationality among the agents.

Education, experience and wsages. Recent trends and long-term evolution, by Dominique Goux et Éric Maurin. This study uses an Employment survey panel (1990-1994) to endeavour to measure wage disparities linked to professional experience and qualifications. To do this, it neutralizes the biases connected with the endogenousness of unobserved individual characteristics as well as those related to the endogenousness of specific employer wage practices. It moreover uses a pseudo-survey panel on training and vocational qualifications to evaluate the development of the wage return on qualifications from 1970 to 1993 by neutralizing the effects of growing unemployment selectivity.

The Econometric Modelling of Individual Transitions on the Labour Market, by Jean-Pierre Florens, Denis Fougère, Thierry Kamionka and Michel Mouchart. This paper presents recent econometric models of individual mobility on the labour market. The general framework is given by the theory of point processes, which are continuous-time finite state space processes. We keep the mathematical formalism at the lowest possible level and specific attention is paid to the main basic concepts. The first part of the paper is devoted to specification problems, but the inference of the introduced models is not addressed in depth. Some structural models leading to point processes are presented and some applications for the French labour market are summarized. The last part of the paper considers the introduction of unobservable heterogeneity and endogenous sampling in specification and inference analysis.