JEL: D24, L33

Article history: Received 8 March 2007 Received in revised form 14 October 2009 Accepted 20 October 2009 Available online xxxx JEL classification: D24 L63 L86

This paper presents an econometric mediation analysis. It considers identification of production functions and the sources of output effects (treatment effects) from experimental interventions when some inputs are mismeasured and others are entirely omitted. JEL Code: D24, C21, C43, C38.

When scientists choose research projects with the highest expected value an externality can appear; slight variations on existing work will be selected in preference to new lines of research that would ultimately generate more value. New research enjoys two advantages: it is riskier and hence more likely to lead to high-value follow-up projects and it can generate more follow-up projects. Less ...

When scientists choose research projects with the highest expected value an externality can appear; slight variations on existing work will be selected in preference to new lines of research that would ultimately generate more value. New research enjoys two advantages: it is riskier and hence more likely to lead to high-value spin-o¤ projects and it can generate more spin-o¤ projects. Less comm...

We propose a model to study the role of industry-level external economies scale in open economies. If elasticity governing strength is below inverse trade each industry, then specialization under frictionless consistent with comparative advantage, tractable even frictions, and all countries gain from trade. External lower gains except if country specializes industries high economies, they ampli...