Training and Age-Biased Technical Change : Evidence from French Micro Data

We model and test the hypothesis that technical and organizational change may be biased againstolder workers. This may occur through a direct adverse effect on their productivity, or throughinsufficient training responses to change. We show that the impact of technical and organizationalchange on the optimal training profile and on the age of retirement is theoretically indeterminate.Using a French matched employer-employee data set, we find evidence that computerized firmsselect their older workers more. But modern firms also tend to train all their workers more,regardless of age. Technical change may thus explain a decline in the employment of older workersthrough a direct unfavorable impact on their productivity rather than through a comparativedisadvantage with regard to training.