![Sample our Urban Studies journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/110826/TOC-Subject-Sample-2021-Urban-Studies.jpg"})
ABSTRACT
This paper argues that urban-regional income development depends on a larger fabric of economic relations at the national and international levels. Focusing on Core-Based Statistical Areas (CBSAs) in the US, the paper identifies firms’ subsidiary networks across space and their changes over time. These networks form a basic architecture through which important growth impulses in production and innovation are transmitted that impact urban income levels. Using a balanced panel of U.S. CBSAs with LexisNexis Corporate Affiliations data from 1993 until 2017, we develop a model that examines the relationship between national and international connectivity and urban income levels, differentiated by origin/destination of ties, industrial sectors, and various interaction effects. Our results strongly support that linkages at both the national and international scale (particularly linkages with European locations) are significantly related to urban-regional income development.
Acknowledgements
Parts of this paper, to which both authors contributed equally, were presented in 2021 at the Annual Meeting of the American Association of Geographers, the Annual Meeting of the Association of Collegiate Schools of Planning, and the Regions in Recovery E-Festival of the Regional Studies Association. The authors would like to thank Giulio Buciuni, Riccardo Crescenzi, Bart Los and Kerstin Schaefer for encouraging feedback at these conferences and Gilles Duranton, He (Shawn) Shuang, Michael Storper, and Yihan Wang for thoughtful comments and suggestions on earlier drafts of this paper.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 While we are broadly interested in regional economic development, we use income levels as a specific indicator throughout our analysis, as this is a useful indicator of people’s economic well-being.
2 Research on clusters has similarly emphasized the role of trans-local connections and ‘global pipelines’, suggesting that it is critical to look beyond the regional boundaries of clusters to understand why these local communities are able to produce continuous growth (Bathelt, Malmberg, and Maskell Citation2004; Giuliani, Balland, and Matta Citation2019).
3 The LexisNexis database includes hundreds of thousands of subsidiary linkages for larger U.S.-based firms and their changes on a year-by-year basis. Subsidiary linkages are typically associated with significant employment numbers although precise data are not available. For instance, retail trade subsidiaries do not include small retail outlets but primarily consist of larger facilities, such as warehouse, logistics, and administrative locations. Linkages are heavily concentrated within the manufacturing sector (see Online Appendix Table A.1), likely associated with substantial economic effects.
4 FDI and connectivity measures are closely related as FDIs generate connectivity for firms and when accumulated to the city level, this is an indicator of the ability of cities to access economic resources elsewhere.
5 Additionally, investments may lead to relational upgrading in value chains and networks that can stimulate regional economic development (Glückler and Panitz Citation2016).
6 It should be noted that we used the number of subsidiary linkages rather than the investment value in our connectivity measures to emphasize opportunities from market access and knowledge flows and avoid bias toward large-scale mining and manufacturing projects.
7 These measures of connectivity are unweighted. However, our main results in Tables 2 and 3 are similar when using a weighted measure of the total volume of linkages rather than unique linkages.
8 A limitation of the QCEW is that the data may be suppressed to protect confidentiality when the total number of employees in an industry and/or a particular area is small. This implies a larger impact at finer levels of industry aggregation, as well as in MicroSAs where employment totals are much smaller than in the MSAs. However, because data suppression only occurs when there is a relatively limited number of observations, we do not expect that there will be substantial measurement error. The fact that the coefficients for High-Tech Share and Manufacturing Share are similar in MicroSAs and MSAs creates confidence that our models capture the impact of these variables on urban-regional income.
9 Transforming the logged variables using the inverse hyperbolic sine transformation (which is defined for zero and has an interpretation as an approximate elasticity) as a robustness check, as well as using the logged share of the population in high-tech, manufacturing, with college degrees, and that is foreign born, both produced very similar results to our main findings in Table 2 column 2.
10 Manufacturing was defined as including the 2-digit SIC codes 20-39. Advanced services were defined as including the 2 digit codes for Communications (SIC code 48), Finance, Insurance and Real Estate (60-67), Business Services (73 but excluding 734: Services to Dwellings and Other Buildings, 735: Miscellaneous Equipment Rental and Leasing, and 736: Personnel Supply Services), Motion Pictures (78 but excluding 783: Motion Picture Theaters and 784: Video Tape Rental), Legal Services (81), and Engineering, Accounting, Research, Management, and Related Services (87).