Abstract: April 1999 - Estimated age and job seniority profiles of wages and marginal productivity in Ethiopia suggest that both skill formation and job matching significantly affect growth of wages and productivity over time. However, job matching is by far the more important of the two sources of growth in wages and productivity. Mengistae analyzes production and labor market data for a random selection of small to medium-size firms in Ethiopia to answer two questions: Does a worker's marginal productivity increase with time in the labor market or with job seniority, as must be the case if on-the-job skill formation or job matching has anything to do with the dynamics of wages observed in the data? Assuming that marginal productivity grows with experience or seniority, is skill formation more or less important than job matching as a source of growth in productivity? The main feature of Mengistae's analysis is the joint regression of the log of the average product of hours in a firm and the log of average hourly earnings of a firm's employees on the shares of experience-seniority cells of workers in total annual hours in the firm. Marginal productivity falls as experience in the labor market passes the 15-year mark, but the expected marginal product of a mobile worker with 16 or more years of experience is still nearly 80 percent higher than that of the base group. The between-jobs growth of hourly wages with potential experience is also large, but not as large as growth in marginal productivity for workers with less than 15 years of experience. Mengistae concludes that job matching is far more important than skill formation as a source of growth in productivity. Net mobility gains account for at least twice the share of the return to skill formation in the observed between-jobs growth of wages with market experience. The rate of return to skills formation is higher in the United States than in Ethiopia. The relative return to skills formation is probably lower in Ethiopia partly because the flow of information about the labor market is more restricted there. This paper-a product of the Development Research Group-is part of a larger effort in the group to identify firm-level sources of growth in productivity. The author may be contacted at [email protected].