Learning-by-Doing Spillovers in the Semiconductor Industry

Abstract

The semiconductor industry is often cited as a "strategic" industry in part because important learning-by-doing spillovers may justify special industrial policies. Documenting the precise nature of these spillovers is crucial for determining the advisability of such policies and is helpful for understanding the contribution of learning to endogenous growth. Yet existing empirical evidence on learning by doing in semiconductor production is scant and evidence on spillovers is nonexistent. Using quarterly, firm-level data on seven generations of dynamic random access memory (DRAM) semiconductors over 1974-92, we find that (a) learning rates average 20 percent, (b) firms learn three times more from an additional unit of their own cumulative production than from an additional unit of another firm's cumulative production, (c) learning spills over just as much between firms in different countries as between firms within a given country, (d) Japanese firms are indistinguishable from others in learning speed, and (e) intergenerational learning spillovers are weak, being marginally significant in only two of seven DRAM generations.

Learning or lock-in: Optimal technology policies to support mitigation

“…This approach is in more detail explained in Fischer and Newell (2007) and relies on the learning-by-doing dynamics elaborated in Spence (1984) and Ghemawat and Spence (1985). It is consistent with econometric studies on external learning-by-doing spillovers which suggest that learning does not only depend on the individual firm's cumulative production but also -to some extent -on the other firms' cumulative output (Irwin and Klenow, 1994;Barrios and Strobl, 2004). From a social planner's perspective, spillovers are irrelevant as cumulative output determines learning.…”

“…learning spillovers and risk premiums, seems to be difficult to quantify. Several econometric studies about learning-by-doing spillovers in manufacturing and semiconductor industry suggest 0.2 ≤ φ ≤ 0.6 (Irwin and Klenow, 1994;Gruber, 1998;Barrios and Strobl, 2004). 11 Within related integrated assessment or policy assessment models, spillover rates usually range between 50 and 80 percent (Jones and Williams, 1998;Popp, 2006;Fischer and Newell, 2008).…”

Energy, the Environment, and Technological Change

“…In the economic literature, these models focus on experience within a single production process, such as the falling costs of production for Liberty ships during World War II (Rapping, 1965) or successive generations of silicon chips (Irwin and Klenow, 1994). In contrast, most of the learning curve studies using energy data make use of more aggregate industry-level data.…”

From tradition to modernity: Economic growth in a small world

“…Similarly, Jaffe et al (1993) find localization effects for the links between patent creation and patent citation on the level of country, state, and metropolitan area, and that localization fades gradually over time. For DRAM production in fully developed countries Irwin and Klenow (1994) show evidence for international knowledge spillovers. At the same time, Foster and Rosenzweig (1995) present evidence that farmers in LDCs are still learning mostly from their neighbors.…”