��ɫ��Ӱ

Note: This paper has been superseded by SIEPR Discussion Paper 15-005.

The first version of this paper was presented to the Conference on R&D, Education and Productivity, held in Memory of Zvi Griliches (1930-1999) on 25-27th, August 2003 at Carrè des Sciences, Ministère de la Recherche, Paris, France. Gabriel Goddard furnished characteristically swift and accurate assistance with the simulations (and the graphics based upon them) that appear here (in Section 4). I am grateful to Wesley Cohen for his perceptive discussion of the conference version. In revising it for publication I have benefited also from the comments of Bronwyn Hall (particularly those clarifying several points regarding Griliches’ redirection of his empirical research toward aspects of technological change and productivity other than diffusion), and from the many suggestions for restructuring the paper that I received from two anonymous referees. Subsequent presentations of the material to the MERIT Workshop on Information Technology and New Industry and Labour Market Dynamics (held on 3-4 June 2004) at the University of Maastricht, and at the seminar of the Eindhoven Centre for Innovation Studies (on 17 November 2004) at Eindhoven University of Technology, provided stimuli to make still other improvements. Later conversations with H. Peyton Young on the general class “heterogeneous threshold” models of adoption have contributed to the paper’s present exposition of the challenges posed to empirical identification of alternative diffusion mechanisms. Lastly, but not least, the wise counsel and extraordinary patience of the editors, Jacques Mairesse and Manuel Trajtenberg, has been vital in bringing the paper to its present state. I am indebted to all those who have offered their help and insist that none should be held to blame for whatever defects and deficiencies remain. This revised version replaces SIEPR Discussion Paper 09-016.