Contenu du sommaire : Analyse des comportements économiques à partir de données de panel

Analyse des comportements économiques à partir de données de panel

• Présentation générale - Patrick Sevestre, Pierre Blanchard p. 1-9
• Effet du pouvoir de négociation et du marché local du travail sur la détermination des salaires - Jennifer C. Smith p. 1-17
  Bargaining Power and Labour Market Influences on Wages Determination by Jennifer C. Smith This paper uses a unique panel of data at the level of the bargaining group to examine aspects of "right-to-manage" models of wage determination. Empirical measures of firms' and unions' bargaining power are identified and found to be important influences on wage setting. The role of union characteristics in wage determination is examined. Results confirm their importance and illuminate previous survey findings. Features of the local labour market are shown to affect bargained wages over and above the influence of aggregate factors.
• Partage explicite ou implicite du profit dans la détermination des salaires - Tim Jenkinson, Sandeep Bhargava p. 19-29
  Explicit Versus Implicit Profit Sharing and the Determination of Wages by Sandeep Bhargava and Tim Jenkinson This paper considers the claim that explicit profit sharing reduces the marginal cost of labour. This is contrasted with the view that implicit profit sharing occurs through wage bargaining. Using a microeconomic data set from the U.K. we find no evidence that the introduction of profit sharing reduces base wages and, hence, the marginal cost of labour. However, firm profitability is found to have a positive effect on wages, which supports the hypothesis of implicite profit sharing through wage bargaining. These findings suggest that it is hard to justify the favourable tax treatment of profit-related pay found in the U.K.
• L'impact de l'allocation chômage sur la probabilité de réemploi : de nouveaux résultats pour le Royaume Uni - Élena Stancanelli p. 31-44
  Unemployment Duration and the Duration of Entitlement to Unemployment Benefit by Elena Stancanelli The aim of this paper is to investigate the impact of the level and the expected duration of employment benefits on the individual probability of leaving unemployment. The relation between the expected duration of unemployment benefit and the probability of leaving unemployment has recently attracted considerable attention. Still a limited number of empirical studies is available for European countries. The data used for the analysis relate to the UK. The probability of leaving unemployment to take up a full-time job is distinguished from the probability of exiting in other states. I conclude that although there is some evidence of spikes in the hazard rate near the time of expected benefit exhaustion, no statistically significant benefit exhaustion effect is found.
• Le marché du travail et l'évolution des salaires dans le transport aérien aux États-Unis : l'effet de la déréglementation de 1978 - Pierre-Yves Crémieux p. 45-61
  The Effect of the 1978 Deregulation on the Air Transport Labour Market and Wages in the United States by Pierre- Yves Crémieux This study compares observed earnings with'simulated values to show that the deregulation of air traffic in 1978 had very different effects on pilots, engineers and flight attendants. Deregulation had no statistically significant effect on engineers' wages. However, the representative flight attendant's earnings decreased by at least 12% from 1978 to 1985 ajnd by at least 39% from 1978 to 1992 compared with what they would have been had there been no deregulation. Pilots' earnihgSjwere 12% lower in 1985 and 22% less in 1992 than they would have been if air traffic had not be deregulated. Engineers' earhijlgs* were more resistant due to the fact that their skills can be more easily transferred and the rriose firnited rents tha't they had accumulated before deregulation. When the effects of deregulation are combined with the effects of the 1982 American recession and the changes in the industrial relations environrnent to the United States to 1980s, the deviation between the predicted values and the simulated values widens signification for the engineers and the pilots, but remains the same for the flight attendants.
• Progrès technique et croissance de la productivité : estimations sur un panel incomplet de firmes ayant des qualités de production différentes - Robert Gagné, Georges Dionne p. 63-76
  Measuring Technical Change and Productivity Growth with Varying Output Qualities and Incomplete Panel Data par Georges Dionne and Robert Gagné This paper presents a method for estimating technical change using incomplete panel data. In addition, we show how estimates of cost elasticities with respect to output qualities can be used to evaluate productivity growth adjusted for quality changes. In order to control for the incomplete nature of the sample used, specific dummy variables for entry in and exit from the sample are introduced into the cost function. The proposed methodology is applied to an unbalanced sample of trucking firms from Quebec. Results show that without correction for sample entry or exit, the estimated rate of technical change between 1987 and 1988 is -29,7%. When entry in or exit from the sample is taken into account, the estimated rate is -4,2%. Also, during the same period, the observed TFP growth rate is 63.4%. After adjustment for quality changes, this rate falls to -4,6%.
• Le ralentissement de la productivité des entreprises d'électricité au Texas : le rôle des marges, des rendements d'échelle et du progrès technique - Subal C. Kumbhakar p. 77-89
  An Anatomy of the Productivity Slowdown: Markups, Returns to Scale, and Technical Change in Electric Utilities in Texas By Subal C. Kumbhakar This paper considers a profit maximizing model to measure total factor productivity (TFP) growth and decompose it into components attributed to technical change, returns to scale and markups. Technical change is further decomposed into pure, nonneutral, and scale augmenting components. The role of these factors is analyzed in accounting productivity growth of electric utilities in Texas. An interesting feature of the study is the use of panel data on individual electric utilities in Texas during 1966-1985, which covers periods before and after the introduction of statewide regulation. The availability of panel data allows us to model markup behavior over time without imposing any a priori behavior on them. The model also introduces and controls for heterogeneity in the cost structure of the utilities. Empirically we find that markups and TFP growth declined substantially during the year of statewide regulation. In spite of this declining tendancy it is found that markups were the major contributing factor in TFP growth during the entire sample period.
• La productivité de la recherche et développement des entreprises industrielles aux États-Unis et en France - Bronwyn H. Hall, Jacques Mairesse p. 91-110
  Estimating the Productivity of Research and Development: an Exploration of GMM Methods Using Data on French and United States Manufacturing Firms Jacques Mairesse and Bronwyn H. Hall A comparative study of the contribution of R&D to firm-level productivity in French and United States manufacturing firms in the 1980s is presentend. The study suses two large panels of approximately 1000 manufacturing firms covering over half of all R&D spending in each country and focuses on the estimation and interpretation of the relationship between output growth and the growth of R&D investment in the prsence of simultaneity and firm heterogeneity. We use GMM methods to control for both sources of estimation bias, and we finf : overail, the contribution of R&D to sales productivity growth appears to have declined during the 1980s, the role of simultaneity bias is higher in the U.S. than in France, possibly reflecting the grater importance of liquidity constraints for R&D investment in that country.
• Estimations de lois de consommation sur un pseudo-panel d'enquêtes de l'Insee ( 1979,1984,1989) - François Gardes, Nilton Cardoso p. 111-125
  Estimating Consumption Laws on a Pseudo-Panel of INSEE Surveys (1979, 1984 and 1989) by Nilton Cardoso and François Gardes Complete demand systems are generally estimated using aggregate time series due to the non-availability of panel data. These estimates are subject to aggregation biases when the specifications are nonlinear in the variables, to instable parameters over a long period and to the inability to estimate sub-population behaviour laws. We construct a pseudo-panel from three Insee family budget surveys by grouping households into homogenous cells to compare their averages on three different dates in accordance with Deaton's methodology (1985). We suggest technical improvements to deal with heteroscedasticity and the cell definition method. Two-thirds of the cross-sectional income elasticities estimated by a quadratic AI equation differ from the first-difference or within dimension estimated time series elasticities. This discrepancy may reflect the endogeny of the individual effects, which bias the cross-sectional elasticities. The social distribution of consumption, such as it is estimated by comparing rich and poor household consumption structures in a survey, can therefore not be used indiscriminately to forecast changes in consumption over time.
• Estimation des élasticités de court et de long termes de la demande d'électricité sur données de panel à partir d'estimateurs à rétrécisseur - Robert P. Trost, G.S. Maddala, Hongyi Li p. 127-141
  Estimation of Short Run and Long Run Elasticities of Energy Demand from Panel Data Using Shrinkage Estimators by Hongyi Li, G.S. Maddala and Robert P. Trost In the analysis of panel data, there are, broadly speaking, three approaches: to present separate estimates of the parameters in each cross -section, present pooled estimates with or without cross-section specific effects, or present a pooled estimator assuming some, but not complete, heterogeneity, as in the random coefficient model. In the last case, we would be interested in the mean of the coefficients for the cross-sections, or the separate coefficients themselves. The paper presents in a unified framework the estimators in this last case, using both the classical, empirical Bayes, and Bayes methods. The methods are illustrated with estimation of elasticities of residential demand for electricity and natural gas in the U.S.
• Estimation de relations de long terme sur données de panel : nouveaux résultats - Alain Pirotte p. 143-161
  Estimation of Long Run Effects on Panel Data: New Results by Alain Pirotte This article is focus on estimated long run effects on panel data. We show that the probability limit of between estimator of a static model, whereas the true specification is a dynamic relationship, gives the long run effects. If the exogenous variable follows a stationary process with individual effects, we show that the convergence of Between estimator is more pronounced than the OLS and Within estimators. Moreover, with individual effects the association OLS and Between estimators with long run effects is proved. Finally, the estimation of a labour demand model on a french panel data of 93 firms over the period 1987 to 1991 confirms the theoretical results.
• Tests de causalité sur données de panel : une application à l'étude de la causalité entre l'investissement et la croissance - Diana Weinhold p. 163-175
  Causaliy Testing in Panel Data with an Application to the Question of Investment and Growth by Diana Weinhold This paper develops an approach for panel causality analysis which allows for flexibility in the causal relationship from unity to unity, allowing estimation of distribution of causality in a possibly heterogeneous panel. Two possible estimation techniques for "mixed fixed and random" coefficients model are explored. These methods both allow the coefficients of an orthogonalized causal variable to vary randomly while avoiding many of the problems associated with random coefficients and dynamics in panels. Using simulations to develop a method for interpreting the estimated distributions, it is then possible to predict the probabilities of causality associated with the panel. This approach is then applied to the question of causality between investment and growth in a panel of countries. The paper find that there is great instability and feedback among countries in this relationship. Therefore, it is clear that the restricting assumptions of traditional pooling models can be inappropriate, especially in simple models. Because the analysis signals this problem, the paper's results further underline the need for the interpretive flexibility that this proposed causality test imparts.
• Utilisation jointe de données de panel et d'informations agrégées dans l'estimation d'un modèle Probit - Michael Lechner, François Laisney p. 177-189
  Combining Panel Data and Macro Information for the Estimation of a Panel Probit Model by François Laisney and Michael Lechner When studying particular subgroups of a population the econometrician typically has few observations, and should draw upon any additional relevant information. We illustrate, for the estimation of a labour-market participation model for lone mothers, the relative benefits derived from using the panel structure of the data in a GMM framework and from including macro information in the form of extra moments.
• Estimation de modèles non linéaires sur données de panel par la méthode des moments généralisés - Michael Lechner, Jörg Breitung p. 191-203
  GMM Estimation of Nonlinear Models on Panel Data by Jôrg Breitung and Michael Lechner We show that the Generalized Method of Moments (GMM) methodology is a useful tool to obtain the asymptotic properties of some existing estimators for non-linear panel data models as well as to construct new ones. Many non-linear panel data models imply conditional moments, which do not depend on parameters from the off-diagonal part of the intertemporal covariance matrix of the error terms. Methods based on these moments sacrifice some efficiency compared to FTML but are much easier to compute since they do not require multivariate integration. The pooled maximum likelihood estimator, the sequential ML estimator based on minimum distance estimation in the second step, and previously suggested alternative GMM estimators are based on these moments. Although the pooled ML estimator is asymptotically the least efficient of the estimators considered, the Monte Carlo study indicates that it may have good small sample properties. We use a low dimensional approximation of the optimal instrument matrix to obtain an estimator, which appeared to be nearly as efficient as FTML. However, GMM estimators are easier to compute and also posses desirable properties.