Subways and the Diffusion of Knowledge: Evidence from China

• Published date: 01 July 2022
• Author: Qiao Wang,Xiuyan Liu,Fan Zhang,Tao Hu
• Date: 01 July 2022
• DOI: http://doi.org/10.1111/cwe.12428

©2022 Institute of World Economics and Politics, Chinese Academy of Social Sciences

60 China & World Economy / 60–99, Vol. 30, No. 4, 2022

*Qiao Wang, PhD Candidate, School of Economics and Management, Southeast University, China. Email:

wangqiao_seu@163.com; Xiuyan Liu (corresponding author), Professor, School of Economics and

Management, Southeast University, China. Email: lxiuyan320@seu.edu.cn; Fan Zhang, PhD Candidate,

School of Economics and Management, Southeast University, China. Email: zhangfanzf_1991@126.com;

Tao Hu, Assistant Professor, The College of Arts and Sciences, Oklahoma State University, the US. Email:

tao.hu@okstate.edu.

Subways and the Diffusion of Knowledge:

Evidence from China

Qiao Wang, Xiuyan Liu, Fan Zhang, Tao Hu*

Abstract

This paper aims to examine the role of the subway network, a typical form of

transportation within urban areas, on innovation and knowledge diffusion. Applying

the difference-in-differences strategy and spatial analysis, we used 1,332 newly opened

stations in China from 2000 to 2013 as a quasi-experiment to identify the local effects

of subway expansions. Results suggested that by reducing communication costs and

increasing opportunities for interaction, subway construction would bring growth at the

district level. Knowledge dissemination would become more active after new stations

open. Micro-level results showed that these positive impacts were highly localized; that

is, only those fi rms located within 1km around stations benefi ted from the new subway.

Moreover, new subways facilitated the flow of knowledge from station to station and

assisted fi rms in acquiring knowledge from more distant technology clusters conveniently.

Keywords: commuting, knowledge diffusion, spatial analysis, subways

JEL codes: O30, R30, R41

I. Introduction

Public transportation systems, like most infrastructure, have a direct impact on

commuting effi ciency and a long-term impact on economic growth, which has prompted

substantial debate regarding their local implications (Gibbons et al., 2019; Yao et al.,

2019; Gendron-Carrier et al., 2022). Intuitively, improvements in transportation facilities

could increase inter-region trade. Evidence from the Raj railroad and Shinkansen

show that this positive effect was not only refl ected in the reduction of trade costs but

also in the expansion of opportunities for trade (Donaldson, 2018). Another result

©2022 Institute of World Economics and Politics, Chinese Academy of Social Sciences

61

Subways and the Diffusion of Knowledge

of the construction of cross-regional traffi c projects is that cities in the network have

received more development opportunities. Railways have increased population mobility

and improved immigration opportunities by reducing travel costs and accelerating

urbanization and economic growth (Duranton and Turner, 2012). Expressways have

affected commuting efficiency within metropolitan areas and have accelerated the

separation between work and residential areas, which may trigger suburbanization

(Baum-Snow, 2007). Besides the flows of trade and immigration, an improvement

in commuting conditions can also facilitate the spread of knowledge across regions

(Agrawal et al., 2017). However, previous research has focused more on the effect of

cross-regional infrastructure. Considering the views of Henderson et al. (2021) on the

locality of technology diffusion and the role of traffic conditions in the connections

between technology clusters within the city (Duranton and Kerr, 2015), the research in

this paper is more interested in the relationship between intra-city transportation and

knowledge diffusion.

Subways are widely used as public transportation facilities in metropolitan areas.

Many studies have been conducted on the economic impact of the subway system,

covering demographics, productivity, housing prices, land value, pollution, and

congestion (Redding and Turner, 2015; Gonzalez-Navarro and Turner, 2018; Li et al.,

2019). Although these studies have used several different rigorous methods, including

regression discontinuity design, difference-in-differences (DID), and structural

estimation to identify systematically the externalities of subways (e.g., Billings, 2011;

Heblich et al., 2020), the relationship between subway construction and innovation

within the urban area still needs to be analyzed. Knowledge spillover depends on ideas

exchanged through communication. Face-to-face communication helps to transmit

information better, especially tacit knowledge, and the spillover effect will attenuate

with the increase in geographic distance (Jaffe et al., 1993; Howells, 2002; Kloosterman,

2008). The development of rapid transportation within the city has shortened the relative

distance between communicators by reducing travel time and increasing the probability

of meeting. The construction of the subway system may therefore have a positive impact

on knowledge spillover. This paper seeks to verify this logic and, through a series of

mechanism analyses, to explain why the subway will impact knowledge fl ows. As far as

the authors are aware, this is the fi rst paper to discuss this issue.

Empirically, for a mature subway system, endogenous problems caused by the

improvement of other infrastructure in the target area will increase the potential risk of

estimation bias and pseudo-causal effects. However, China’s subway system provides

a quasi-experiment. Using Chinese cases, we therefore used the time-varying DID

approach to evaluate the impact of subway expansions on knowledge diffusion. Our

©2022 Institute of World Economics and Politics, Chinese Academy of Social Sciences

62 Qiao Wang et al. / 60–99, Vol. 30, No. 4, 2022

empirical analysis suggests a significant positive impact of subway expansions on

local knowledge diffusion. Specifi cally, at the district level, new subways have brought

growth and stimulated corporate innovation, which has caused an increase in all kinds

of patents. Such a positive effect is not only reflected in innovation but also in the

fl ow of knowledge. After the opening of new stations, the citation of local patents has

become more frequent. However, this effect is heterogeneous in terms of economic

density and industrial development policies. We found that subway system expansions

in regions with lower density caused a greater increase in knowledge diffusion, and the

impact was concentrated in high-diversification and low-specialization districts. The

fi rm-level analysis suggested that the scope of local effect on knowledge spillovers had

been limited. The positive spillover effect captured by the spatial specification could

only be identifi ed within a 1km radius, implying a highly localized effect on knowledge

diffusion. Moreover, such an effect was stronger when the new stations were located far

from the central business district (CBD) and research and development (R&D) institutes,

and firms with large sizes gained more benefits than small ones. More importantly,

we found that travel time reduction from the subway made firms obtain knowledge

from more distant areas within the city. Compared with the previous subways, the new

subway extended the distance between patents with citation relations by 80.1 percent

within the city, and the patents applied by fi rms located near the new stations had more

opportunities to be cited.

This study makes several contributions. First, it extends research on the impact

of transportation improvement on innovation and knowledge diffusion from the inter-

city level to the intra-city level, which may interest urban and economic geography

scholars. Second, it fi nds that the growth effect of innovation and knowledge diffusion

at the district level has been more related to the highly localized growth around the

new stations than the spillover effect within a larger buffer. This is a new fi nding that is

different from previous studies. Third, the study provides new and direct evidence for

the high correlation between the knowledge fl ow trajectory and transportation facilities

within the city. The improvement of urban transportation facilities has extended the

distance of knowledge dissemination and has made inventors obtain new knowledge

from more distant knowledge clusters. Finally, the results regarding the heterogeneity

effect have important policy implications for urban managers. Our research provides a

basis for constructing subways in low-density areas or locations far from the innovation

core within the city.

In the rest of the paper, Section II reviews related literature and summarizes the

mechanism by which subways affect the knowledge fl ows from previous studies. Section

III briefly reviews the history of subway construction in China. Our identification