Ubiquitous digital technologies and spatial structure; an update

Abstract

This paper examines the impact of widespread adoption of information and communication technologies (ICT) on urban structure worldwide. Has it offset agglomeration benefits and led to more dispersed spatial structures, or has it strengthened urban externalities and thus resulted in more concentrated spatial structures? Theoretical and empirical studies on this question have produced contradictory findings. The present study recognizes that assumptions made earlier about the evolution of technological capabilities do not necessarily hold today. As cutting-edge digital technologies have matured considerably, a fresh look at this question is called for. The paper addresses this issue by means of several data sets using instrumental variable methods. One is the UN data on Urban Settlements with more than 300, 000 inhabitants. Estimation methods with these data show that increased adoption of ICT has resulted in national urban systems that are less uniform in terms of city sizes and are characterized by higher population concentrations in larger cities, when concentration is proxied the Pareto (Zipf) coefficient for national city size distributions. Two, is disaggregated data for the urban systems of the US, defined as Micropolitan and Metropolitan Areas, and for the UK, defined as Built-up Areas in England and Wales, respectively. These data allow for the impacts to be studied for cities smaller than those included in the cross-country data. Increased internet usage improved a city’s ranking in the US urban system. Similarly, increased download speed improves a built-up area’s ranking in England and Wales.

Can the internet reshape the national city size distribution? Cross‐country evidence

“…Our finding is different from the decentralization effect found by Ioannides et al (2008), Tranos and Ioannides (2015, 2020) and from the concentration effect found by Tranos and Ioannides (2021). The reason for this could be attributed to the difference between their sample countries and ours in the development phases of the internet.…”

“…At the inflection point, the fixed broadband penetration rate is 18%, and the internet user rate is 45%, which is calculated based on 2.5 users per household. The percentage of internet users is less than 45% in almost all sample countries of Ioannides et al (2008) and in most sample countries of Tranos and Ioannides (2015, 2020), while that in most sample countries of Tranos and Ioannides (2021) is more than 50%. Almost all of sample countries used by Ioannides et al (2008), Tranos and Ioannides (2015, 2020) are on the left side of the inflection point in our study, and most sample countries used by Tranos and Ioannides (2021) are on the right side of the inflection point in our study.…”

“…First, surprisingly scarce empirical studies have focused on this topic. Only four studies have examined the effects of the internet on national city size distribution (Ioannides et al, 2008; Tranos & Ioannides, 2015, 2020, 2021), but their results are not consistent. Second, the city population data used for measuring national city size distribution in the above cross‐country studies are mainly from the censuses of each country and are defined in various ways; thus, the data are less comparable.…”

“…Using city population data from the same source as Ioannides et al (2008) but over different time spans, Tranos and Ioannides (2020) estimated a WLS regression and reached the same conclusion as Ioannides et al (2008). However, using annual population data from the United Nations for urban agglomerations with 300,000 inhabitants or more in 38 countries, Tranos and Ioannides (2021) adopted a country fixed‐effects model and conducted a 2SLS analysis to verify the effect of the internet on city size distributions but obtained results opposite to those of Ioannides et al (2008). In addition to these studies, Tranos and Ioannides (2015) also used city population data for 2014 from the City Population Project, but this study encompassed a wider range of countries and creatively analysed national heterogeneity.…”

“…Actually, the population data used in these studies are incomparable between countries. Moreover, only two of these papers (Tranos & Ioannides, 2020, 2021) addressed the potential endogeneity caused by the unobservable heterogeneity between countries by using a panel fixed‐effects model or weighted least square (WLS) regression; however, their samples covered only 38 and 24 countries, respectively. These insufficiencies might lead to biased research results and a misunderstanding of the effect of the internet on city size distribution, resulting in inappropriate spatial planning policies and a misallocation of resources.…”