Learning by Doing and Technology Sharing in Asymmetric Duopolies

The existence of an experience curve for the firm has theoretical foundation in Arrow’s pioneering work (Arrow, 1962). The study of the learning-curve effect was initiated, however, almost thirty years earlier by Wright (1936). It was further developed by Alchian (1959), Hirschleifer (1962), Hirschmann (1964), and by an important work by the Boston Consulting Group (1972). The learning-curve phenomenon appears to be related, in the first place, to the introduction of new products. It has been observed that a doubling in accumulated production of a new product can result in a decline of unit costs by a factor of 10 to 50 percent of their initial level (Teng and Thompson, 1983). Empirical documentation concerning the learning curve can be found in several papers that appeared in the period from the early sixties to more recent years (e.g., Lundberg, 1961; Alchian, 1963; Hollander 1965; Zimmerman, 1982; Lieberman, 1984; Alder and Clark, 1991). This literature considers cost reduction as an outcome of experience and cumulative output is the variable commonly chosen to represent that experience. The revival of interest in the study of the learning curve can be attributed to some important recent contributions that suggest the existence of a link between learning by firms and incremental innovations in both production processes and products (see in particular Rosenberg, 1976; Nelson and Winter, 1982).