THE COMPETITIVE SEMICONDUCTOR MANUFACTURING HUMAN RESOURCES PROJECT:

Second Interim Report
CSM-32
Clair Brown, Editor

10. How does Knowledge Flow? Inter-Firm Patterns in the Semiconductor Industry *
Melissa M. Appleyard **

10.1 Introduction
10.2 Why Would Rivals Share?
10.3 Cross-Industry Comparison of Knowledge-Sharing: Semiconductors vs. Steel
10.4 Knowledge-Sharing in Different Institutional Contexts: Japan vs. the U.S.
10.5 Survey Results
10.6 Summary and Conclusions
REF References

Abstract

When considering the evolution of technology, economists often assume the existence of knowledge spillovers, especially across company boundaries. The vast majority of the spillovers literature is silent on the following points: How does knowledge flow across company boundaries? How do industry characteristics and national institutions shape knowledge diffusions? To what degree do companies direct knowledge flows? This research provides a foundation for answering these questions by examining knowledge-sharing patterns in the semiconductor industry. An analysis of survey data shows that public sources of technical information play a larger role in knowledge diffusion in Japan relative to the U.S. and in the semiconductor industry relative to the steel industry. Only through an understanding of knowledge flows can company managers and policy makers establish optimal patterns of knowledge diffusion.

10.1 Introduction

This paper documents the existence of inter-firm information flows in a knowledge-intensive industry. Although many economists assume that knowledge "spills" across company boundaries, this research details the actual mechanisms by which technical knowledge is disseminated in the semiconductor industry. This research employs the concept of knowledge-sharing, defined as the transfer of useful know-how or information across company lines and executed through communication channels that vary both in terms of their privacy and their legal status. Although previous empirical studies have documented informal inter-firm knowledge exchanges, this paper explores the reasons why patterns differ across industries and across countries. The findings from survey data offer a rare account of the primary knowledge-sharing vehicles in the semiconductor industry, both in the U.S. and Japan, and provide a grounding for optimal strategies of knowledge management.

Section 10.2 provides a framework for thinking about the payoffs to knowledge-sharing. This section begins by distinguishing between the access to knowledge and its use. Access to knowledge is either public or private in nature, and its subsequent use may or may not be legally restricted. Based on a typology of findings from empirical studies of private, unrestricted inter-firm knowledge-sharing, the conditions under which firms, even rival firms, would exchange knowledge are presented.

Turning to a cross-industry analysis, Section 10.3 compares the characteristics of the semiconductor industry to those of the steel industry. I hypothesize that industries with a rapid pace of technological change, such as the semiconductor industry, would have a lower incidence of private knowledge-sharing.

By focusing on distinctions between the institutional environments in the U.S. and Japan, Section 10.4 discusses how such distinctions might shape knowledge-sharing within a given industry. Taking into account the influence of the employment systems and the intellectual property right regimes in the two countries, I predict that private knowledge-sharing would be more likely among U.S. semiconductor workers than among their Japanese counterparts. This prediction stems from a higher incidence of employee turnover in the U.S. and America's first-to-invent patent system.

Drawing on surveys of people in the semiconductor industry, Section 10.5 presents empirical evidence supporting the two hypotheses posited above and provides clues to the profiles of people who engage in knowledge-sharing. In contrast to previous empirical studies of the steel industry, public channels of communication play a central role in the transfer of knowledge in the semiconductor industry. As for the international comparison, U.S. semiconductor workers are approached more frequently for technical information and are more likely to fulfill at least one request per year. However, Japanese workers are more likely to fulfill the majority of requests that they receive. The field work in Section 10.5 details the primary mechanisms of knowledge-sharing in the semiconductor industry.

The final section summarizes the findings and describes how a company can incorporate knowledge-sharing patterns into its knowledge management strategy. The final section also discusses policy implications derived from the evidence presented in this paper.

Go to Next Section of this Chapter
Go to Table of Contents for this Chapter
Go to Table of Contents for the CSM-HR Interim Report

* Presented at the Tenth Annual Franco American Seminar on the Economics and Econometrics of Innovation Strasbourg, France, June 1996.

** Department of Economics, UC Berkeley (email: appleyar@econ.berkeley.edu). I gratefully acknowledge the Alfred P. Sloan Foundation for funding. I am indebted to Sarah Bales, Clair Brown, Bronwyn Hall, Rose Marie Ham, David Mowery, Steve Raphael, Paul Romer, Stephan Schrader, my dissertation group, the participants in UC Berkeley's Institutions Seminar, the participants in the ADRES Strasbourg Conference on Innovation and many people associated with the semiconductor industry, including Tony Alvarez, Jose Arreola, Neil Berglund, Dan Flamm, Jim Nulty, John Schuler, and the late Vinay Sohoni for useful comments and insights.