Working Paper: CEPR ID: DP3734
Authors: Ralph Siebert
Abstract: In this study we derive a structural econometric model of learning by doing with multiproduct competition from a dynamic oligopoly game. We show the importance to account for multiproduction effects through product differentiation when measuring learning by doing. Using quarterly firm-level data for the dynamic random access memory semiconductor industry, we provide evidence that accounting for multiproduction results in lower learning effects and firms behaving more competitive in the product market. We can confirm that firms follow intertemporal production plans for investing in future cost reductions. We also find that learning effects are higher at the beginning of the life cycle.
Keywords: dynamic random access memory; dynamics; economies of scale; learning by doing; multiproduct firms; product life cycle; product market competition; semiconductors; spillovers
JEL Codes: C10; L10; L60; O30
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| multiproduction effects (E23) | lower learning effects (C92) |
| multiproduction effects (E23) | more competitive behavior (L13) |
| lower learning effects (C92) | lower output (E23) |
| more competitive behavior (L13) | firms price close to marginal costs (D40) |