![Sample our Built Environment journals, sign in here to start your access, Latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5926/TOC-Subject-Sample-2021-Built-Environment.jpg"})
Abstract
Breschi S. and Lenzi C. The role of external linkages and gatekeepers for the renewal and expansion of US cities’ knowledge base, 1990–2004, Regional Studies. This paper examines the role of external linkages and gatekeepers for the renewal and expansion of cities' knowledge base in US metropolitan co-invention networks. It is argued that the relative importance of direct external linkages and external relations mediated by gatekeepers varies according to specific local conditions. It is found that direct external relations, on average, contribute to broadening and rejuvenating the local knowledge base and outperform external links mediated by gatekeepers; the latter, however, are especially important in cities with a localized and specialized knowledge base, as they enable the trans-coding and absorption at the local level of externally sourced knowledge.
Breschi S. and Lenzi C. 外部连结与把关人在美国城市知识基础的更新与扩张中的角色，1990 年至 2004 年，区域研究。本文检视美国大都会的共同创新网络中，城市知识基础更新与扩张的外部连结与把关人的角色。本文主张，直接的外部连结和把关人中介的外部关係的相对重要性，随着特定的在地条件而异。本研究发现，平均而言，直接外部关係导致在地知识基础的扩张与復苏，并胜过由把关人中介的外在连结；但把关人在具有在地化且专殊化的知识基础的城市中却特别重要，因为他们能将外来知识进行代码转换，并促进地方层级的吸收。
Breschi S. et Lenzi C. Le rôle des liens externes et des gardiens quant au renouveau et à l'expansion de la base de connaissance des grandes villes aux E-U, de 1990 à 2004, Regional Studies. Cet article cherche à examiner le rôle dans les réseaux de coinvention métropolitains aux E-U des liens externes et des gardiens quant au renouveau et à l'expansion de la base de connaissance des grandes villes. On affirme que l'importance relative des liens externes directs et des relations externes facilités par des gardiens varient selon les conditions locales spécifiques. En moyenne, il s'avère que les relations externes contribuent à l’élargissement et au rajeunissement de la base de connaissance locale et font de meilleure figure que les liens externes facilités par des gardiens. Cependant, ces derniers revêtent une importance particulière dans les grandes villes dont la base de connaissance est localisée et spécialisée, parce qu'ils permettent de convertir et d'assimiler au niveau local la connaissance recherchée à l'extérieur.
Breschi S. und Lenzi C. Die Rolle von externen Verknüpfungen und Gatekeepern für die Erneuerung und Erweiterung der Wissensbasis in Städten der USA, 1990–2004, Regional Studies. In diesem Beitrag wird die Rolle von externen Verknüpfungen und Gatekeepern für die Erneuerung und Erweiterung der Wissensbasis von Städten der USA in metropolitanen Netzwerken für gemeinsame Erfindungen untersucht. Wir argumentieren, dass die relative Wichtigkeit von direkten externen Verknüpfungen und von durch Gatekeeper vermittelten Außenbeziehungen je nach den spezifischen lokalen Bedingungen unterschiedlich ausfällt. Es stellt sich heraus, dass direkte Außenbeziehungen im Durchschnitt zur Erweiterung und Verjüngung der lokalen Wissensbasis beitragen und leistungsfähiger sind als durch Gatekeeper vermittelte externe Verknüpfungen; allerdings sind diese in Städten mit einer lokalisierten und spezialisierten Wissensbasis besonders wichtig, da sie die Transcodierung und Absorption auf der lokalen Ebene des extern beschafften Wissens ermöglichen.
Breschi S. y Lenzi C. Papel de los vínculos externos y los guardianes para la renovación y la expansión de la base de conocimiento en las ciudades estadounidenses, 1990–2004, Regional Studies. En este artículo analizamos el papel de los vínculos externos y guardianes para la renovación y la expansión de la base de conocimiento en las ciudades estadounidenses en redes de coinvención metropolitana. Sostenemos que la importancia relativa de los vínculos externos directos y las relaciones externas mediadas por los guardianes varía en función de las condiciones locales específicas. Observamos que las relaciones externas directas contribuyen en general a ampliar y rejuvenecer la base de conocimiento local y superan a los vínculos externos mediados por los guardianes; sin embargo, estos últimos son especialmente importantes en las ciudades con una base de conocimiento localizada y especializada porque permiten la transcodificación y absorción en un ámbito local del conocimiento de origen externo.
Notes
1. MSAs are defined by the US Office of Management and Budget (OMB) as urban core areas of at least 50 000 people, plus adjacent counties that have a high degree of social and economic integration with the core, as measured by commuting ties.
2. The US Patent and Trademark Office (USPTO) patent data might appear a more natural choice for a study on invention in US cities. However, the average quality of USPTO patents has declined considerably in the period under consideration due to a series of concomitant factors. Because of these trends, patents of insufficient quality or with inadequate search of prior art were issued more often (Hall et al., Citation2004; Jaffe and Lerner, Citation2004). The use of EPO data reduces this noise as only more valuable patents are extended to Europe. Patents are dated by using their priority year, i.e. the first date at which the patent was applied for anywhere in the world as this date is closest to the actual time of the invention.
3. Fuller details on the routine implemented for data cleaning and standardization are provided by Lissoni et al. (Citation2006).
4. The June 2003 definition of MSAs issued by OMB was used. For details and maps, see http://www.census.gov/population/www/metroareas/metrodef.html/.
5. Note that the adjectives ‘internal’ and ‘external’ used to characterize the network ties only refer to the spatial boundaries delimiting a city. Thus, a co-invention tie between two inventors located within the same city is internal to that city, whereas a co-invention tie between two inventors located in different cities is external to them, regardless of the organizational affiliation of the two inventors. In particular, in the case of firms with plants located in different cities, the external reach indicator considers also those links that may arise between inventor i working in city c for applicant a and all inventors working for applicant a in every other US city. Although one may question that such types of links are more likely to be established (and possibly mediated by firm-level practices and strategies), from a spatial perspective they do not differ from the links established by inventor i working in city c for applicant a with all the other inventors in any other US city and working for other firms than a.
6. For disconnected (i.e. not reachable) pairs of inventors, and, therefore,
is equal to
, which reduces to 0.
7 Precisely, for pairs of inventors (ij), where ,
is set equal to 0. Namely, it was assumed that no knowledge flow is taking place for a distance above that threshold.
8. To be more specific, consider two technological classes i and j in which a city's patents before time t have been classified, but which were not combined together in any patent. Suppose that a patent at time t is co-classified in i and j. This patent introduces a new combination of technologies. Yet, this recombination does not expand or renew the existing knowledge base of a city. For this reason, it is not included in the computation of new pairs of technological codes.
9. Groups are next divided into subgroups; however, subgroups are nested into groups (i.e. their hierarchical level varies across groups) and therefore cannot be exploited in a study such as this one. For further details, see http://www.wipo.int/export/sites/www/classifications/ipc/en/guide/guide_ipc.pdf/.
10. An illustrative case of the problems that may arise from the inclusion of these outliers is that of patent EP1632470, which has been applied for by NPS Pharmaceuticals, located in Salt Lake City, Utah. This patent, with priority year 1991, is classified in 40 different technological groups. Of the 780 combinations of technological groups associated to this patent, 702 were new to Salt Lake City in 1991. Given that in the same year this company applied for other three patents classified in a similarly large number of different technological groups, the total number of new combinations introduced in Salt Lake City in 1991 was 2978, which is an order of magnitude higher than the number of new combinations in 1990 (210) and 1992 (212).
11. The authors also tried to include in the analysis the total number of technological groups in which the internal patents made in a city in the period (t − 1, t − 5) are classified. Yet, we were forced to exclude it because of problems of multicollinearity with the other independent variables. The authors thank two anonymous referee for raising this issue.
12. Although the adoption of a relatively aggregated technological classification level may not allow the details about the technological content of patent data to be exploited fully, the measurement of this variable is based on technological subclasses (i.e. four-digit IPC) and not technological groups (i.e. the lower technological aggregation level), as the use of groups would disproportionately and artificially inflate its value.
13. More formally, Pi
is defined as:where Pdi
is the number of patents that city d has made in technological field I; and wdc
is the weight of city d on all external collaborative links between inventors of city c and inventors in all other cities. This implies that cities with which inventors of city c have more collaborative links (and thus exchange knowledge) weight more in the computation of the Krugman index of city c. Similarly, only cities with which inventors of city
have external linkages enter into the computation of the index.
14. In affiliation networks as co-invention networks, this coefficient tends to increase with the number of inventors per patent (Newman, Citation2010). In fact, any three inventors listed in the same patent document by construction form a triangle, i.e. all paths of length two are automatically closed. By extension, the greater the number of patents with more than two inventors, the greater the coefficient. To correct for this possible bias and inflationary mechanism, the generalized version of the clustering coefficient proposed by Opsahl (Citation2013) was computed, which excludes those triples of inventors connected because of a joint patent and only counts the number of triples (i.e. closed paths of length two) that are the outcome of independent interactions between pairs of inventors.
15. Diagnostic tests tend to exclude serious risks of collinearity. The variance inflation ratio (VIF) is below 2.5 for all variables, with the exception of the (log) of internal patents (3.52), the share of external reach mediated by the gatekeepers (3.12), and the Krugman index (2.56). However, even for these variables, VIF is well below 10, which is the rule of thumb value usually considered in order to detect serious problems of multicollinearity.
16. For a unit change in , the expected count in the dependent variable changes by a factor of
, holding all other variables constant. For a standard deviation change in
, the expected count changes by a factor of
. Alternatively, IRR can be interpreted in terms of percentage change in the expected count. Thus, for a unit change in
, the percentage change in the expected count is given by
.
17. Note that is the IRR associated with the average external reach.
18. The two values correspond, respectively, to the 25th and 75th percentiles of the empirical distribution of the variable. In this way, one can simulate what happens to the expected number of new combinations when the share of external reach mediated by the gatekeepers increase from small to large values.
19. See note 7.
20. See note 7.
21. To put it in a different way, the shortest path from f to U does not pass through g. More generally, there are cases in which an inventor reaches an external inventor through both direct connections and indirect links, mediated by other inventors. As indirect links imply, by definition, a greater distance than direct links, these triples are obviously excluded from the computation of the external reach mediated by gatekeepers. This happens particularly in the case of patents with more than two inventors, one of whom is located outside the boundaries of the city.