Report on the 1st IZA European Summer School in Labor Economics, 1998

Logo

The two lectures of the 1st IZA summer school were held by Francis Kramarz (CREST/INSEE, Paris) on the "Analysis of Labor Markets Using Matched Employer-Employee Data", and Frank Windmeijer (Institute for Fiscal Studies, IFS, London) on the "Evaluation of Job Training Programs".
In his six lectures, Prof. Kramarz gave an extensive introduction into the specific estimation techniques and problems connected with the use of matched employer-employee data. First, Prof. Kramarz presented a basic econometric model in which some measured outcome for an individual is a function of time-varying exogenous characteristics of the individual, a pure person effect, a pure firm effect, and a statistical residual. Prof. Kramarz then used this basic econometric model to present a whole range of questions within labor economics which can be stressed using matched employer-employee data, i.e. the relative importance of person and firm variables in the determination of compensation, the relative importance of individual mobility in relation to firm-specific employment adjustments, unemployment insurance, and other aspects of the labor market. For instance, using a matched employer-employee data from France he showed, that person effects explain about 90 % of inter-industry wage differentials and about 75 % of the firm-size wage effect, while firm effects explain relatively little of either differential. In the final part of his lecture, Prof. Kramarz described a broad list of existing employer-employee datasets and discussed with the students further research questions which could be answered with these datasets.

 

The lectures of Prof. Frank Windmeijer were about the evaluation of training programs, in particular the typical econometric problems evolving in this area of research and the possible strategies to solve these problems. The first part of the lecture was dedicated the endogeneity bias problem and instrumental variables. Within this part Prof. Windmeijer discussed several instrumental variable estimators together with some specification test procedures, the generalized method of moments (GMM), as well as a dummy endogenous variable model. Constructing counterfactuals is the central problem in the literature on evaluating social programs. Therefore, Prof. Windmeijer gave in the second part of his lecture a detailed description of the evaluation problem and presented the prototypical solutions to this problem, i.e. the Before-After Estimator, the Difference-in Differences Estimator and the Cross-Section Estimator. Prof. Windmeijer then talked about the advantages and problems of Social Experiments. The last part of his lecture was dedicated to the various sample selection problems evolving in the empirical research on training programs and possible solutions to this problem (instrumental variables estimator, classical selection model, matching estimator, different panel estimators).