I report new evidence for localized knowledge spillovers identified by within-patent variations in the geographic matching rates of citations added by inventors and citations added by examiners. Evaluated at the mean citation lag, inventor citations are 20 percent more likely than examiner citations to match the country of origin of their citing patent, while US inventor citations are 25 percent more likely to match the state or metropolitan area of their citing patent. The localization of intranational knowledge spillovers declines with the passage of time, but international borders present a persistent barrier to spillovers.

Several theories of firm performance can explain the well known observation that survival is positively related to age. However, a more mundane explanation—selection bias driven by variations in firm quality—may also underlie the phenomenon. This paper employs a 90 year plant-level panel data set on the U.S. iron and steel shipbuilding industry of the nineteenth and early twentieth centuries to discriminate between the explanations. The shipbuilding industry exhibits the usual joint dependence of survival on age and size, but this dependence is eliminated after controlling for heterogeneity by using preentry experience as a proxy for firm quality. The evidence points to a dominant role for selection bias in creating the age dependence of survival. At the same time, preentry experience is found to have a large and extremely persistent effect on survival, and this finding is inconsistent with standard explanations for the role of preentry experience on firm performance.

A new data set facilitates study of learning spillovers in World War II shipbuilding. Our results contain two principal but contrasting themes. First, learning spillovers were a significant source of productivity growth, and may have contributed more than conventional learning effects. Second, the size of the learning externalities across yards, as measured by Spence's θ, were small. These findings, which are not mutually inconsistent, suggest an optimistic view of learning spillovers: they are a significant source of productivity growth, but the market failures induced by learning externalities may be modest. (JEL D24, N72, O3)

This paper offers some new estimates of the contribution of learning to the rapid increases in labor productivity observed in the construction of Liberty ships during World War II. The study exploits new data on physical capital investment and vessel quality constructed from contemporary records held at the National Archives. Estimates of the rate of learning are shown to be sensitive to the inclusion of the new capital data, and data on vessel quality provide evidence that part of the measured productivity increases were secured at the expense of quality.

Over the last twenty years, many studies have been made of the elasticity of substitution between capital and labor. The reported estimates are highly variable, and reveal an apparent dichotomy between cross-sectional and time-series studies. The former suggest that capital and energy are substitutes while the latter suggest the converse. All these studies reported Allen partial elasticities of substitution. We suggest that the Morishima elasticity may be a more useful measure for the issues of concern to capital energy substitution. We calculate the Morishima elasticities from parameters estimated in a selection of earlier studies and find no excessive variability, nor any evidence of the time-series/cross-section dichotomy. Capital and energy are Morishima substitutes. Copyright 1995 by MIT Press.

Patent Citations and the Geography of Knowledge Spillovers: A Reassessment by Peter Thompson and Melanie Fox-Kean. Published in volume 95, issue 1, pages 450-460 of American Economic Review, March 2005

Patent Citations and the Geography of Knowledge Spillovers: A Reassessment: Reply by Peter Thompson and Melanie Fox-Kean. Published in volume 95, issue 1, pages 465-466 of American Economic Review, March 2005

Using a large, individual-level wage data set, we examine the impact of a major technological innovation—the steam engine—on the demand for skills in the merchant shipping industry. We find that the technical change created a new demand for engineers, a skilled occupation. It had a deskilling effect on production work—moderately skilled able-bodied seamen were replaced by unskilled engine room operatives. On the other hand, able-bodied seamen, carpenters, and mates employed on steam vessels earned a premium relative to their counterparts on sail vessels, and this appears partly related to skill.