یادگیری از تجربه یا یادگیری از دیگران؟ استنتاج آموزش غیر رسمی از یک تابع سود سرمایه انسانی با داده منطبق با کارفرما، کارمند

A model of informal training which combines learning from own experience and learning from others is proposed in this paper. It yields a closed-form solution that revises Mincer–Jovanovic's [Mincer, J., Jovanovic, B., 1981. Labor mobility and wages. In: Rosen, S. (Ed.), Studies in Labor Markets. Chicago University Press, Chicago, pp. 21–64] treatment of tenure in the human capital earnings function. We estimate the structural parameters of this non-linear model on a large French cross-section with matched employer–employee data. We find that workers on average can learn from others 10% of their own human capital on entering one plant, and catch half of their learning from others’ potential in just 2 years. The private marginal returns to education are declining with education as more educated workers have less to learn from others and share the social returns of their own education with their less qualified co-workers. The potential for learning from others on the job varies across jobs and establishments, and this provides a new distinction between imitation jobs and experience jobs. Workers in imitation jobs, who learn most from others, tend to have considerably longer tenure than workers in experience jobs. Although workers in experience jobs can learn little from others, we find that they learn a lot by themselves. We document several analogies between the imitation jobs/experience jobs “dualism” and the primary/secondary jobs and firms’ dualism implied by the dual labor market theory. However, our binary classification of jobs depicts the data more closely than the dual theory categorization into primary-type and secondary-type establishments. Competition prevails between jobs and firms but jobs differ by their learning technology.

The effects of human capital on earnings are commonly captured by a remarkably simple equation, which was suggested and estimated by Mincer (1974) on US census data and is still known as the “Mincerian” earnings function. The most widely estimated version of this model is linear in education and quadratic in labor market experience: it is usually called the quadratic earnings function. An extended version of this equation, which was proposed by Mincer and Jovanovic (1981), also includes a quadratic function of tenure in the incumbent firm. The Mincerian earnings function has been efficient in extracting valuable information about the costs and returns of education and training from experience-earnings profiles. The recent availability of large matched employer–employee data sets in a number of countries (Abowd and Kramarz, 1999) makes it worth asking how this popular tool could be extended to extract additional information about the amount and structure of informal learning on-the-job. A natural direction of research was advocated by Mincer (1974) himself: “[…] the most important and urgent task is to refine the specification of the post-school investment category […] to include details (variables) on a number of forms of investment in human capital.” As a matter of fact, matched worker–firm data yield valuable information about co-workers and firms’ training policy which makes it possible to separate learning from others and learning from own experience. This should contribute to a better understanding of the processes of human capital accumulation used by firms and of firms’ heterogeneity in this respect. We shall be using here a unique French cross-section on labor cost and wages structure (INSEE 1992) comprising 150,000 wage earners in 16,000 establishments. Much of the informal training taking place on-the-job may be captured by a combination of learning from own experience (or, self-learning) and learning from others (or, learning by watching). Barron et al. (1989) confirm the importance of these informal learning processes in the US. In the 3 months following the recruitment of new workers, 96% of on-the-job training is given to them in an informal way by other workers (145.2 h of a total 151.1 h) and more than one-third of on-the-job training (53.1 h) is provided through a “learning by watching” process. Learning by oneself through experience and learning from others seem to capture the essential ingredients of informal learning on-the-job, so that a model that incorporates these two elements should offer a good description of informal on-the-job training. They both form the microeconomic counterparts of the autonomous and catch-up growth processes separated by Benhabib and Spiegel (1994) in macroeconomic growth models, following a suggestion of Nelson and Phelps (1966). The model of informal on-the-job learning from self and others (LSO) presented in Section 2 has the nice feature of yielding a closed-form solution.1 Hence, it is possible to identify its structural parameters with matched employer–employee data and infer from the latter the characteristics of informal training within jobs and firms, like the relative use of both learning technologies, the complementarity or substitutability of education and training, the magnitude of spillover effects of human capital among co-workers and social rates of return to education. The data and econometric approach are then discussed in Section 3. Section 4 compares the econometric estimates of the non-linear LSO earnings equation with those derived from Mincerian polynomials in tenure (linear, quadratic or quartic) and interprets the results. Since the potential for learning from others is a novel distinguishing feature between jobs and establishments in our analysis, Section 5 examines the heterogeneity of jobs and establishments along this new dimension. This exercise provides a new partition of jobs by their learning technology: learning from own experience in experience jobs and learning from others in imitation jobs. These two types of jobs are contrasted and the correspondence between the imitation jobs/experience jobs “dualism” and the primary/secondary jobs and firms’ dualism implied by dual labor market theory ( Doeringer and Piore, 1971 and Dickens and Lang, 1985) is questioned. The summary of findings and conclusions will be found in Section 6.

We have suggested a simple model of informal learning on-the-job which combines learning from (own) experience and learning from others. This yields a closed-form solution that revises the Mincer–Jovanovic's (1981) treatment of tenure in the human capital earnings function by relating earnings to the individual's job-specific learning potential. We estimated the structural parameters of this non-linear model on a large French cross-section with matched employer–employee data. We find that workers on average can learn from others 10% of their own human capital on entering the firm, and catch half of their learning potential in just 2 years. Since individuals learn fast from their co-workers, the estimated returns to tenure loom larger than predicted by a quadratic, or even a quartic-in-tenure, Mincerian function in the first years and decline more sharply (until about 30 years). Learning by watching accounts for three quarters of the marginal rate of return in the first year of tenure, but this share falls rapidly, with an average of 12%. While education and self-learning on-the-job are complementary, education and learning from others on-the-job are substitutes. The more education, the less can be learned from others. This forces the private marginal return curve to decline with education, an effect which was not captured by current theory. Seen from a different perspective, the more educated workers share the social returns of their own education with their less qualified co-workers. The potential for learning from others on the job varies across jobs and establishments, and this provides a new distinction between imitation jobs and experience jobs. Workers in imitation jobs, who learn most from others, tend to have considerably longer tenure than workers in experience jobs. The latter are more mobile and have accumulated more market experience. Although workers in experience jobs can learn little from others, we find that they learn a lot by themselves. Consequently, we do not find a close correspondence between the imitation jobs/experience jobs “dualism” and the primary/secondary jobs and firms’ dualism implied by the dual labor market theory. Even though imitation jobs imply far less turnover than experience jobs, imitation jobs do not appear to be “better” in terms of education levels and wages. We show, however, that predictions of the dual labor market theory which cannot be observed at the job's level under our classification of jobs emerge from the aggregation of jobs at the establishment level. Furthermore, we find no evidence of rationing of primary-type jobs and establishments. Competition prevails between jobs and firms but jobs differ by their learning technology. Firms that make an intensive use of learning from others adhere rather naturally to more collective forms of workers’ governance such as reliance to trade unions in comparison with those that make an intensive use of self-learning.