Effects From Expanding High-Speed Railway Network on Regional Accessibility and Economic Productivity in China

Abstract

This paper reviews the historical evolution of China’s high-speed railway (HSR) network, evaluates improvements in accessibility, and analyzes the associations of improved accessibility with regional economic productivity, given regional heterogeneity. Three accessibility indicators were considered: average travel time to all prefectural level regions (ATT), average travel time to important cities (ATI), and daily accessible prefectural level regions. These indicators were used to quantify the development of the HSR network during two periods, from 2007 to 2012 and 2012 to 2018. First, the results revealed that, in the first period, the accessibility indicators of the east region improved the most, whereas in the second period, the west significantly improved. We subsequently analyzed the economic productivity and -equilibrium of the urban agglomerations affected by introducing the HSR. Those results suggested that the Triangle of Central China, Chengdu-Chongqing, and Central Guizhou urban agglomerations performed well as the HSR developed. The linkages between regional economic productivity and accessibility improvement were then measured using a multivariable regression with panel data. The results showed that the reduction of ATT and ATI significantly positively contributed to economic productivity at different geographical scales in China. Furthermore, ATT had a larger effect in the northeast and central regions, whereas ATI had a larger effect on the northeast and west regions.

Keywords

rail passenger rail transportation high speed sustainability and resilience transportation and society transportation in the developing countries policy and planning transportation and sustainability transportation and economic development economic impacts

The high-speed railway (HSR) network was an important technological breakthrough in the development of passenger transportation in the 20th century. Since the opening of the Shinkansen, an HSR in Japan, several countries have gradually introduced an HSR network. There are currently more than 20 countries with operating HSR networks. As the largest network worldwide, 29,904 km of HSR lines were operating in China by the end of 2018.

The HSR is believed to be a powerful engine for economic growth, owing to its advantages of speed, reduced travel time, improved capacity, and increased accessibility. Several studies have investigated the impact of HSR services on regional economic development; however, these studies have reached significantly different conclusions because of the different research scales, methodologies, and study periods employed. One perspective is that investments in HSR have a significant positive impact on economic growth ( 1 ), increasing resident income ( 2 ). Another perspective is that investments in HSR have a negative effect on economic growth in underdeveloped areas ( 3 ). The industry agglomeration effect created by HSR is also a hot topic, but the direction of agglomeration differs with different types of industry and the different scales of the HSR network ( 4 , 5 ). The heterogeneity issue cannot be ignored when estimating the effect of HSRs.

China has become one of the least equal economies in the world ( 6 ) with regional economic disparities and an urban–rural gap. This could eventually threaten the country’s social stability and economic sustainability. Regions of high population density and higher economic productivity are mainly concentrated in the eastern part of China. At the same time, the Chinese central government has long pursued a biased development of its public infrastructural elements, concentrating its investments in the eastern regions and coastal areas ( 7 ). To achieve a balanced economy and harmonious society, the Chinese central government has launched a series of policies to prioritize the development of underdeveloped areas. With respect to the construction of HSR lines, the government has stressed that HSR services will cover the entire national landmass. HSR network investment is considered a policy tool for narrowing economic inequalities ( 8 ). However, the distribution of factors supporting regional economic productivity, such as capital, education, technology, and information, has been uneven for many years. In China’s medium- and long-term railway planning, balancing the economic development between eastern, central, and western regions is the ultimate goal. In addition, the railway network planning of urban agglomeration is an important part of national planning. With the rapid development of the HSR over the last decade, the urban agglomerations in the eastern region, such as the Yangtze River Delta and Pearl River Delta urban agglomerations, have more mature HSR networks, whereas the urban agglomerations located in the western region, such as along the Yellow River in Ningxia, Central Yunnan, and Central Guizhou, are either without an HSR or have only one HSR line. It is therefore important to conduct a more detailed, multilevel impact assessment of the Chinese HSR development to answer the questions, “What is the impact of improved accessibility created by HSR expansion on regional economic output?” and “Do the impacts differ across regions?” Conducting a valid impact assessment of the HSR could help answer those questions and provide direction and guidance for future HSR network expansion.

There are two categories of research on the impacts of HSRs on regional economies. The first category of research employs econometric models to analyze the economic growth created by HSR development and focuses on the flow of production factors generated by HSRs, including labor, capital, technology, and information. This approach typically considers agglomerations and spillover effects. For instance, Chen et al. employed a computable general equilibrium (CGE) model to investigate the impact of HSR investment on the economy and environment at national level ( 9 ). The existing research, which employs a CGE model, is based on time series data, making it difficult to identify effects on low spatial scales ( 1 ). The multiple regression models were widely adopted for panel data sets. Wetwitoo and Kato empirically analyzed the impact of the Japanese HSR on regional productivity, using an ordinary least squares estimation model, fixed-effects model, and instrument variable model ( 8 ). Heuermann and Schmieder used a gravity model to estimate the effect on workers’ mobility caused by HSR expansion in Germany ( 10 ). Shao et al. investigated the influence of the HSR on urban service industry agglomeration in the Yangtze River Delta region by using a difference-in-differences model ( 4 ). Jin et al. used a spatial economic model to examine the aggregate growth effect and potential spatial spillover effects of HSRs ( 11 ). The second category of research includes studies that focus on changes in accessibility and evaluates spatial and temporal patterns on national or regional scales. For example, Spiekermann and Wegener measured the accessibility of the European transport network, and classified European cities into those that economically outperformed and underperformed based on the per capita gross domestic product (GDP) relative to the potential accessibility index ( 12 ). Shaw et al. divided China’s HSR development into four stages to evaluate the accessibility patterns of cities based on timetable data ( 13 ). Cao et al. examined 49 cities to analyze the impact of HSRs on the distribution of accessibility ( 14 ). Chen and Haynes focused on provinces and municipalities as the basic research units to measure accessibility improvement by HSR development and estimate the economic impact ( 15 ). Jiao et al. applied weighted average travel time, daily accessibility, and potential accessibility to evaluate the accessibility spatial patterns of prefecture-level cities ( 16 ). However, most studies of how HSR development has improved accessibility and affected economics in China were concentrated around the years 2012 and 2014. Given the sharp increase of HSRs since then, these conclusions should be updated.

With respect to accessibility, some researchers have explored the equity issue of accessibility improvements across regions (15, 17) and have claimed that inequalities have increased in China (16, 18). In Spain, completing the HSR did not lead to the overall growth of the national economy, and rapid growth was generally seen in major urban agglomerations. In contrast, the growth in medium and small cities along the HSR line was found to be significantly smaller ( 19 ). This is because the developed regions were competitive in attracting business and capital; as such, their economy was likely to be enhanced by improved accessibility through HSR development ( 19 ). Similar results were found with several European HSR systems, as accessibility in major cities was most improved (relative to peripheral cities) after the introduction of HSR services ( 20 ). In the case of Japan, Sasaki et al. evaluated the impact of the Shinkansen HSR and found that expanding the Shinkansen system had a strong positive effect on developed regions ( 21 ). In contrast, the contribution to less developed regions was more limited. Similarly, Kim and Sultana found that cities near the already advantaged Seoul capital area significantly benefited from having the HSR ( 22 ).

As China’s railway network is the largest and most complex in the world, accessing the economic impact at different geographic scales for national, regional, and urban agglomerations is an impressive undertaking. This study divided the Chinese mainland into east, central, west, and northeast regions, and examined 19 urban agglomerations to assess the different effects of the HSR on accessibility and economics. Most previous studies have explored the economic effect brought about by HSRs based on a dummy variable indicating whether there is an HSR station in the city (1, 3). In our research, we first calculated the interregion travel time before and after the HSR development, based on railway timetables. Second, three accessibility indictors for each region in the years of 2007, 2012, and 2018 were calculated. Finally, the linkage between the HSR and economic productivity was explored, based on an accessibility indictor. The goals of this study were to present the spatial and temporal features of the HSR network expansion from a historical perspective, and to quantify the impact of the HSR construction on accessibility changes from a spatiotemporal perspective. We explored the relationship between accessibility and regional economic productivity at a multilevel geographic scale, while considering regional heterogeneity. The quantitative analysis could provide support for national comprehensive transportation planning.

This paper is organized as follows: the next section introduces the data and research scope and provides descriptive statistics of the data set, followed by descriptions of the methods applied in the study. The expansion of the HSR network is then explained, and the detailed spatiotemporal distribution of accessibility and the associated economics are presented. The associations of regional accessibility with regional per capita GDP are estimated using multiple regressions at different geographic scales. The last section summarizes the conclusions with a discussion about the policy implications.

Data and Research Scope

The research period covers three points: 2007, 2012, and 2018. The HSR was not available in 2007, and only a small portion (9,356 km) was open in 2012. Also, the HSR network with four horizontal and four vertical lines almost entirely opened in 2018, with mileage reaching 29,904 km. Table 1 lists the variables used in this research. All the demographic and economic data for the years 2007 (without an HSR), 2012 (with an emerging HSR), and 2018 (with a well-developed HSR) were collected from the China City Statistical Yearbook for 2008, 2013, and 2019. The missing data were supplemented with data from the statistical bulletin of each prefectural level region in the corresponding years. To avoid overestimating the economic impact of the HSR, we added variables relating to other transportation modes. The daily departure schedules of each airport were collected from the FlightStats website (https://www.flightstats.com), which releases historical flight data including flight number, origin airport, destination airport, and so forth. We collected flight data between July 3 and 9 in 2007, 2012, and 2018, which generated 36,891, 55,865, and 92,612 records, respectively. The descriptive statistics of the variables are shown in Table 2. The coeﬃcient of variation (CV) was adopted to measure the level of regional inequality for each variable.

Table 1.

Definition of Variables

Variables	Definitions	Sources
PGDP	Gross domestic product (GDP) per capita (yuan)	CCSY 2008, 2013, and 2019
INV	Total investment in ﬁxed assets (10,000 yuan)	CCSY 2008, 2013, and 2019
EOST	Expenditure for science and technology (10,000 yuan)	CCSY 2008, 2013, and 2019
SPG	GDP from second industry as a percentage of total GDP (%)	CCSY 2008, 2013, and 2019
NEP	Number of employed persons (10,000 persons)	CCSY 2008, 2013, and 2019
POP	Total population (10,000 persons)	CCSY 2008, 2013, and 2019
AFF	Daily departure flight frequency	https://www.flightstats.com

Note : CCSY refers to the China City Statistical Yearbook.

Table 2.

Descriptive Statistics of Variables

Region	Variable	2007				2012				2018
Region	Variable	Mean	Min.	Max.	CV	Mean	Min.	Max.	CV	Mean	Min.	Max.	CV
Nation (288)	PGDP	21,489.41	3,398	98,938	0.77	42,919.14	8,157	182,680	0.62	61,045.41	12,656	191,942	0.57
	INV	4,538,842.80	410,115	44,586,098	1.25	12,253,045.99	933,488	92,542,052	1.03	21,841,329.16	636,650	186,614,050.9	0.98
	NEP	39.63	4.37	544.38	1.25	52.64	6	856.2	1.57	59.54	6.17	819.30	1.49
	EOST	23,286.95	469	1,057,666	3.85	66,151.30	1,329	2,454,279	3.12	160,268.90	1,231	5,549,817	3.22
	SPG	48.81	15.93	89.59	0.24	51.45	17.1	87.96	0.20	42.84	15.75	72.9	0.22
	POP	423.35	18.14	3,235.32	0.71	441.85	19.8	3,343.4	0.71	455.10	21	3,404	0.72
	AFF	17.68	0	611.71	3.64	26.74	0	869.57	3.38	44.00	0	1,076.57	2.78
East (87)	PGDP	31,295.82	8,424	91,911	0.63	54,629.10	17,425	145,395	0.50	83,813.91	25,367	191,942	0.49
	INV	7,617,439.40	1,096,928	44,586,098	0.97	17,241,585.83	2,241,572	83,402,588	0.83	31,728,688.36	5,180,857	106,433,118.9	0.66
	NEP	61.48	5.49	544.38	1.22	82.50	9.7	717.4	1.28	98.79	14.24	819.30	1.28
	EOST	56,758.43	2534	1,057,666	2.76	150,366.24	6,822	2,454,279	2.35	361,704.77	11,709	5,549,817	2.43
	SPG	51.99	26.83	76.24	0.16	49.67	20.6	70.85	0.16	43.35	18.27	62.21	0.19
	POP	502.76	53.52	1,378.86	0.53	522.88	57.3	1,426.9	0.54	552.56	61	1,462	0.54
	AFF	33.97	0	611.71	3.04	48.68	0	869.57	2.92	75.67	0	1,076.57	2.38
Central (80)	PGDP	16,101.21	5,515	41,917	0.51	34,748.43	12,616	89,903	0.47	53,017.96	21,589	136,920	0.45
	INV	3,284,534.44	633,233	17,327,895	0.92	10,285,928.74	1,255,724	50,160,766	0.80	21,995,794.71	2,149,026	86,458,313.84	0.77
	NEP	32.99	7	159.58	0.67	42.55	7.8	191.9	0.74	47.81	8.45	209.77	0.80
	EOST	9,043.81	894	74,914	1.38	34,058.34	5,014	226,720	1.29	113,650.45	5,616	1,344,072	1.69
	SPG	49.09	23.79	68.77	0.23	54.31	25.2	73.44	0.17	45.80	17.72	63.31	0.17
	POP	452.96	73.65	1,166.72	0.54	482.41	74.2	1,229.2	0.55	492.4	111	1,259	0.54
	AFF	6.87	0	139.29	3.41	10.83	0	176.57	3.18	19.67	0	296	2.89
West (84)	PGDP	16,554.30	3,398	98,938	0.95	36,533.65	8,157	182,680	0.81	51,380.52	12,656	181,486	0.59
	INV	2,819,287.43	437,834	31,615,147	1.59	9,274,646.95	933,488	92,542,052	1.35	17,148,099.71	1,037,916.64	186,614,050.9	1.42
	NEP	25.38	4.37	229.67	1.26	38.22	6	856.2	2.49	41.90	6.17	613.54	1.89
	EOST	7,049.77	469	110,513	1.98	22,589.37	1,329	298,409	1.74	54,108.57	1,231	730,705	2.26
	SPG	46.50	20.14	89.59	0.29	51.45	26.69	87.96	0.24	42.94	19.67	72.9	0.24
	POP	362.30	18.14	3,235.32	1.04	376.94	19.8	3,343.4	1.04	386.33	21	3,404	1.06
	AFF	15.51	0	245.14	2.89	24.66	0	333.86	2.74	42.92	0	501.86	2.49
Northeast (37)	PGDP	21,285.32	7,814	67,161	0.61	47,548.08	18,474	142,067	0.52	46,807.24	23,517	109,550	0.47
	INV	3,915,854.05	410,115	23,618,726	1.28	11,538,233.35	1,203,043	56,253,999	1.13	8,913,593.45	636,650	55,428,857.44	1.33
	NEP	34.99	9.18	148.34	0.79	36.99	9.1	138.3	0.81	32.64	8.87	120.99	0.91
	EOST	12,242.51	925	1,18943	2.07	36,419.95	2,355	392,872	2	28,431.70	1,526	356,622	2.34
	SPG	45.96	15.93	84.97	0.29	49.41	17.1	80.88	0.24	35.03	15.75	54.82	0.29
	POP	311.24	89.51	987.32	0.64	310.97	92.4	993.5	0.65	301.41	78	952	0.66
	AFF	7.69	0	96.57	2.83	14.25	0	153.57	2.47	24.55	0	216.14	2.33

Note : CV = coeﬃcient of variation; Min. = minimum; Max. = maximum. Please see Table 1 for definitions and units of variables; The numbers in parentheses refer to sample size with demographic and economic data for the nation and four economic regions.

Official railway timetables for 2007 (without the HSR) and 2012 (with an emerging HSR) were published by the China Railway Press; timetables for 2018 (with a well-developed HSR) were collected from the Chinese railway’s website. The rail travel time matrix between railway stations in 2007, 2012, and 2018 was obtained using the Floyd shortest path algorithm. The matrices of the 3 years had sizes of 2,243 $\times$ 2,243, 2,527 $\times$ 2,527, and 3,129 $\times$ 3,129, respectively.

Our research sample is 361 geographical units in total, which consist of four municipalities (Beijing, Shanghai, Tianjin, and Chongqing), 333 prefectural level regions, and 24 county-level regions not covered by prefectural level regions. These areas constitute the Chinese mainland. The boundary of each unit was established before 2011, and was slightly changed later. The observed value of our research units is spatial aggregation. We identified the effects brought about by HSR development by measuring interactions among the research units. We investigated the relationship between the units’ economic productivity and interunit railway travel time reductions. The travel time from an origin unit (r₁) to a destination unit (r₂) was calculated as follows:

t = t (r_{1} \to s_{1}) + t (s_{1} \to s_{2}) + t (s_{2} \to r_{2})

(1)

where

$t$ = shortest travel time between the origin and destination,

$t (r_{1} \to s_{1})$ = shortest travel time from the origin to the closest railway station on the highway,

$t (s_{1} \to s_{2})$ = shortest time by railway between two stations, and

$t (s_{2} \to r_{2})$ = shortest travel time between the destination to the closest station on the highway.

If there was at least one station in the prefectural level region, then the travel time between the station and region was considered to be 0, namely $t (r_{1} \to s_{1}) = 0$ and $t (s_{2} \to r_{2}) = 0$ . As such, three 361 $\times$ 361 travel time matrices were obtained for the accessibility analysis. In our study, when calculating interunit travel time, we only considered onboard travel time, ignoring the transfer time and access time to railway stations, only accounting for the reduction in travel time brought about by the HSR development.

The 361 research units were divided into four economic regions; the region’s geographic boundaries are shown in Figure 1a. The eastern region consists of 10 provinces: Hebei, Beijing, Tianjin, Shandong, Shanghai, Jiangsu, Zhejiang, Fujian, Guangdong, and Hainan. The central region includes six provinces: Shanxi, Henan, Anhui, Hubei, Hunan, and Jiangxi. The western region includes the western part of Inner Mongolia, Guangxi, Shaanxi, Gansu, Qinghai, Ningxia, Xinjiang, Chongqing, Sichuan, Guizhou, Yunnan, and Tibet. The northeast region includes Liaoning, Jilin, Heilongjiang, and the eastern part of Inner Mongolia. This classification was initially proposed by the Development Research Center of China’s State Council and has been widely adopted for research and for regional development policy making ( 23 ). Taiwan, Hong Kong, and Macao were not included in this study because the data were not available. Demographic and economic data were available from 288 units out of the 361 regions, which provided the basis for the regression analysis. In 2018, the central government identified 19 unban agglomerations as the economic growth pole. The development degrees, states, and processes of the urban agglomerations are different. The location of each urban agglomeration is shown in Figure 1b. Evaluating the interaction between transportation and economic development in each urban agglomeration could provide policy implications for guiding sustainable development.

Method

Given the spatially aggregated panel data set and different geographic scales, we were interested in understanding the linkages between HSR development and economic productivity at multilevel geographic scales. Generally, HSR development could contribute to a reduction in travel time and to improvements in regional accessibility, ultimately facilitating regional economic productivity. We hypothesized that accessibility improvement would have positive effects on the regional economy. With the substantial regional gaps in China, however, the effects were unlikely to be uniform.

For this study, three accessibility indicators were used to quantify the quality of the HSR network in China; ATT and ATI represent the travel cost indicators of each research unit, with calculations based on 361 $\times$ 361 travel time matrices; ATT(min) is the average travel time to all units, and ATI(min) is the average travel time to the large cities: Beijing, Shanghai, Guangzhou, and Shenzhen. The lower the values of ATT and ATI, the greater the accessibility. The third indicator is a daily accessibility indicator (DA4), which refers to the number of research units that can be reached within 4 h by HSR. The higher the value of DA4, the greater the accessibility. These three indicators have been used extensively in transportation accessibility research (24, 25).

To estimate the impact that HSR development and improved accessibility had on regional economics, we established a panel fixed-effect regression model. The model is specified as Equation 2,

Y_{it} = α + β AA C_{it} + θ X_{it} + μ_{it}

(2)

where $Y_{it}$ represents the natural log of GDP per capita in prefectural level region $i$ at year t, and $AC C_{it}$ represents accessibility-related variables in prefectural level region $i$ at year $t$ . There were three pairs of estimates reflecting the average travel time (natural log of ATT), the average travel time to important cities (natural log of ATI), and daily accessibility. $X$ contains a set of control variables, including the natural log of the fixed-asset investment per capital (lnPINV), natural log of expenditure for science and technology per capita (lnPEOST), the share of production in the secondary industry (SPG), the natural log of the proportion of employed individuals to the total population (lnPNEP), and daily flight frequency (AFF). The disturbance term $μ_{it}$ is the sum of the unobserved region-specific effect $μ_{i}$ and a classical random disturbance $ϑ_{it}$ . Hausman’s test was conducted to assess whether there was a region-specific effect or not. If the test score was significant, the fixed-effect model was preferred; otherwise, the random-effect model was preferred.

Historical Process of HSR Network Development

China’s HSR network comprises new rail lines with a maximum speed of 250 km/h. The network has also upgraded lines with an average design speed of 200 km/h. The construction of China’s HSR was based on the “Medium- and Long-Term Railway Network Plan,” first published in 2004 by the National Development and Reform Commission, Ministry of Transport, and China Railway Corporation. It was revised in 2008 and 2016. The planning of the HSR network has gone through three stages, in response to shifts in its defined purpose. Before 2004, there was only one passenger-dedicated line (Qinhuangdao–Shenyang Passenger Dedicated Line). The main purpose outlined in the first edition of the Plan was to separate passenger and freight. To achieve this, the government approved “4 vertical–4 horizontal” fast passenger-dedicated lines and planned for three intercity passenger HSR transport systems, to address the areas around Bo Hai, the Yangtze River Delta region, and the Pearl River Delta region, respectively. In the second edition of the Plan, the 4 vertical–4 horizontal fast passenger-dedicated lines were slightly adjusted and plans were issued to construct more intercity passenger-dedicated lines for 10 urban agglomerations, including the Triangle of Central China, Chengdu-Chongqing, Central Plains, and the Shandong Peninsula. In 2016, it was acknowledged that the HSR network was unbalanced, with underdevelopment of the west and the northeast regions. As such, in the latest edition of the Plan, the planning HSR network was expanded into an “8 vertical–8 horizontal” network, and included plans for intercity passenger railway networks in 19 urban agglomerations. The objective was to connect all cities with a population over 500,000 by 2030.

The first official HSR (Beijing–Tianjin Intercity Passenger Dedicated Line) was launched in 2008. Figure 2 presents maps of the HSR network from 2008 to 2018 and includes the HSR elements that had been opened by the end of each year. The first HSR line in the west region started operating in 2010; the HSR network was not further expanded until 2014. The HSR connected the Xinjiang–Gansu–Qinghai provinces within the east region; the Guangdong–Guangxi–Guizhou HSR connected the most developed eastern region with the less developed western region in 2014. Liang et al. took counties covered by this typical line as a research sample to evaluate the economic effect on different provinces ( 26 ). In 2017, the HSR connected the Xinjiang–Gansu–Qinghai and Sichuan provinces with the Shannxi province, strengthening HSR connectivity between the western region and others. In 2018, with the exception of a small section from Beijing to Chengde, the 4 vertical–4 horizontal HSR network was opened. Details of that network are shown in Table 3.

Table 3.

The Details of “4 Vertical–4 Horizontal” Fast Passenger Dedicated Lines

Line	Connected cities	Connected regions	Year of full opening
Vertical
1	Beijing–Tianjin–Jinan–Xuzhou–Bengbu–Nanjing–Shanghai	Capital–Central–Southeast region	2011
2	Beijing–Shijiazhuang–Zhengzhou–Wuhan–Changsha–Guangzhou–Shenzhen	Capital–Central–Southeast	2012
3	Beijing–Chengde–Shenyang–Harbin (Dalian)	Capital–Northeast	Most parts were opened in 2012 and 2018
4	Shanghai–Hangzhou–Ningbo–Fuzhou–Shenzhen	Eastern coastal region	2013
Horizontal
1	Xuzhou–Zhengzhou–Lanzhou	East–Central–Northwest	2017
2	Nanjing–Wuhan–Chongqing–Chengdu	East–Central–Southwest	2014
3	Hangzhou–Nanchang–Changsha–Guiyang–Kunming	East–Central–Southwest	2016
4	Qingdao–Jinan–Shijiazhuang–Taiyuan	East–Central	2018

Table 4 shows the distribution of HSR service in each region over time. By 2018, there were 957 HSR stations, covering 68% of the units; 63% of the HSR stations were located in the east and central regions. In 2014, the railway network in Western China had significantly improved. In 2018, more than 80% of the units were covered in the central and east regions, whereas there were less than 50% in the western regions.

Table 4.

High-Speed Railway (HSR) Mileage and Stations Over Time

Year	Length of the HSR (km)	Number of HSR stations					Percentage of covered units
Year	Length of the HSR (km)	Nation	East	Central	West	Northeast	Nation (361) (%)	East (103) (%)	Central (87) (%)	West (130) (%)	Northeast (41) (%)
2008	671.50	29	19	3	0	7	4.71	9.71	2.30	0	12.20
2009	2,698.70	89	49	29	0	11	11.63	18.45	19.54	0	14.63
2010	5,133.40	168	99	42	16	11	19.11	33.98	25.29	4.62	14.63
2011	6,601.00	219	140	47	16	16	23.27	44.66	27.59	4.62	19.51
2012	9,356.00	289	162	68	21	38	30.47	49.51	42.53	5.38	36.59
2013	11,028.00	380	204	87	50	39	36.29	58.25	44.83	13.08	36.59
2014	16,456.00	528	231	140	117	40	48.20	64.08	63.22	29.23	36.59
2015	19,838.00	676	273	173	153	77	56.79	71.84	74.71	34.62	51.22
2016	22,980.00	778	304	216	181	77	60.39	74.76	75.86	41.54	51.22
2017	25,164.00	851	322	234	201	94	63.43	75.73	80.46	44.62	56.10
2018	29,904.00	957	363	238	226	130	68.42	83.50	81.61	47.69	68.29

In the latest version of the “Medium- and Long-Term Plan,” urban agglomerations have been divided into three categories of activity for intercity HSR network planning purposes. The Chinese government anticipated that the HSR would lead to the sustainable development of urban agglomerations. The first category (ID 1–7) involved forming a dense HSR network; the second category (ID 8–12) involved constructing a backbone HSR network; the third category (ID 13–19) involved constructing the backbone HSR line. The number of HSR stations and the HSR coverage rate are shown in Table 5. By the end of 2018, most of the urban agglomerations had a high coverage ratio, with the exception of the urban city groups in the northern slope of the Tianshan Mountain, and along the Yellow River in Ningxia. The percentage of HSR stations in urban agglomerations in 2012 and 2018 was 90% and 84%, respectively. China attached great importance to developing the HSR in urban agglomerations to help to diminish regional inequalities and facilitate economic growth.

Table 5.

High-Speed Railway (HSR) Stations and Coverage Rate of Urban Agglomerations

ID	Name of urban agglomeration	Total units	2012		2018
ID	Name of urban agglomeration	Total units	Number of stations	Coverage rate by HSR (%)	Number of stations	Coverage rate by HSR (%)
1	Jing–Jin–Ji	12	19	66.7	45	91.7
2	Yangtze River Delta	27	61	44.4	114	96.3
3	Pearl River Delta	14	33	64.3	83	92.9
4	Triangle of Central China	31	35	51.6	130	93.5
5	Chengdu–Chongqing	16	13	25	72	75
6	Central Plains	28	17	46.4	41	64.3
7	Shandong Peninsula	8	11	50	38	87.5
8	West Side of the Straits	17	28	47.1	66	94.1
9	Harbin–Changchun	11	13	45.5	59	90.9
10	Mid–Southern Liaoning	10	22	80	45	100
11	Guanzhong Plain	10	4	20	18	60
12	Beibu Gulf	15	3	6.7	30	93.3
13	Central Yunnan	5	0	0	12	100
14	Central Guizhou	6	0	0	27	83.3
15	Northern Slope of Tianshan Mountain	5	0	0	2	20
16	Along the Yellow River in Ningxia	4	0	0	0	0
17	Hohhot–Baotou–Erdos–Yulin	4	0	0	9	75
18	Lanzhou–Xining	9	0	0	14	55.6
19	Central Shanxi	4	1	25	10	75

Regional Accessibility Improvement and Economic Development

Spatiotemporal Pattern of Accessibility Indicator

The spatial distribution of the three accessibility indicators for 2007, 2012, and 2018; and the accessibility improvements between 2007 to 2012 and 2012 to 2018 are shown in Figure 3. Accessibility was classiﬁed into six levels using the Geometrical Interval classification method in ArcGIS. The color change from red to green indicates the changes in the variable as accessibility declined. It shows a core-periphery pattern. In 2007, without the HSR, the ATT, ATI, and DA4 indicated that the most accessible units were Zhengzhou at 853 min, Ganzhou at 512 min, and Shangqiu at 41 units, respectively. In 2018, the most accessible units were Zhengzhou at 499 min, Guangzhou at 217 min, and Kaifeng near Zhengzhou at 113 units. The changes in the accessibility indicators indicated there was a corridor effect in the east, central, and northeast regions between 2007 and 2012. From 2012 to 2018, the units with increasingly shorter travel times were in Yunnan, Guizhou, Guangxi, Sichuan, and Gansu provinces.

Table 6 compares the accessibility indicators between regions in 2007, 2012, and 2018 (before and after the opening of the HSR). The results showed significant improvements in regional accessibility, with the HSR development reducing the average travel time by half. The percentage shown in parentheses is the ratio in the reduction of ATT and ATI and the improvement ratio of DA4 at a national scale and at the four regional scales for the periods 2007 to 2012 and 2012 to 2018. During the first stage (2007 to 2012), the national-level improvements in ATT, ATI, and DA4 were 20.20%, 34.87%, and 78.96%, respectively. The greatest improvements were seen in the east and northeast of China. During the second stage (2012 to 2018), the national rates of improvements in ATT, ATI, and DA4 were 30.05%, 27.42%, and 87.04%, respectively. In this period, the greatest improvements were seen in the west of China. In the first stage, the average travel time to big cities was reduced by over 34.87%, whereas, in the second stage, the average travel time, and daily accessible units of each prefectural level region changed more. At the national scale, the disparity associated with the two travel cost indicators increased. In the central and eastern regions, the difference in ATT kept increasing, whereas ATI increased and then decreased. In the west, the differences in the three indicators continued to increase. In the northeast region, both the differences and fluctuations in indicators were small. Significant changes in accessibility may have significant macro effects on national and regional economies.

Table 6.

Descriptive Statistics of Accessibility Indicators at Different Geographical Scales

	Nation (361)		East (108)		Central (87)		West (130)		Northeast (41)
	Mean	CV	Mean	CV	Mean	CV	Mean	CV	Mean	CV
Descriptive statistics of accessibility indicators in 2007
ATT	1,372.39	0.38	1,300.33	0.23	1,003.91	0.09	1,626.99	0.43	1,528.10	0.15
ATI	1,190.78	0.56	870.34	0.29	729.18	0.20	1,708.22	0.48	1,334.60	0.15
DA4	11.60	0.75	12.63	0.74	18.20	0.48	5.79	0.67	13.42	0.46
Descriptive statistics of accessibility indicators in 2012
ATT	766.15 (30.1%)	0.57	677.68 (29.2%)	0.28	543.15 (30.3%)	0.15	960.17 (32.6%)	0.66	846.45 (21.0%)	0.17
ATT	1,095.22 (20.2%)	0.48	957.18 (26.4%)	0.27	779.00 (22.4%)	0.12	1,423.78 (12.5%)	0.51	1,071.23 (29.9%)	0.16
ATI	775.55 (34.9%)	0.82	492.66 (43.4%)	0.54	389.33 (46.6%)	0.31	1,284.53 (24.8%)	0.62	691.89 (48.2%)	0.24
DA4	20.76 (79.0%)	0.98	25.55 (102.3%)	0.81	37.17 (104.3%)	0.62	6.79 (17.3%)	0.79	18.15 (35.3%)	0.54
Descriptive statistics of accessibility indicators in 2018
ATI	562.86 (27.4%)	0.85	397.56 (19.3%)	0.46	317.95 (18.3%)	0.29	830.32 (35.4%)	0.82	649.81 (6.1%)	0.19
DA4	38.82 (87.0%)	0.81	46.78 (83.1%)	0.56	70.15 (88.7%)	0.44	16.72 (146.1%)	0.94	22.46 (23.8%)	0.48

Note : CV = coeﬃcient of variation; ATT = average travel time to all prefectural level regions (min); ATI = average travel time to important cities (min); DA4 = daily accessible prefectural level regions (units).

The Spatial Pattern of Economic

The descriptive statistics in Table 1 show that the east region was consistently more developed over the study period; the average PGDP value was nearly twice that of the central and west regions. Economic growth in the central region was significant during 2012 to 2018 and exceeded the mean value of the western region’s PGDP. Regional economic disparities, measured in per capita GDP, declined in the period 2007 to 2018. Figure 4, a to c, show the spatial patterns of the economics in China and reveals that the spatial distribution of PGDP in 2007 and 2012 was similar. In 2018, some prefectural level regions in the central region had relatively high levels of per capita GDP. Most developed units are located in coastal areas; however, others have a small population and are rich in mineral and tourism resources, such as units in Inner Mongolia, Qinghai, Gansu, and Xinjiang provinces. The economic growth differed at different stages. Between 2007 and 2012, Inner Mongolia and Guizhou provinces had a high PGDP growth rate. In contrast, the growth rates were high in the second stage for Jiangsu, Anhui, Jiangxi, Hubei, Fujian, Hainan, Chongqing, Sichuan, Yunnan, Guizhou, and Tibet.

Accessibility and Economic Development in Urban Agglomeration

Table 7 provides the accessibility indicators, the GDP per capita (PGDP), and the proportion of the GDP of urban agglomerations to the total GDP of all research units (POG). The CV for PGDP is provided in brackets. The proportion of GDP indicates the strength and importance of urban agglomeration. Table 7 indicates that HSR development led to positive performance by some urban agglomerations; there was an increase in the economic proportions of POG, and the regional differences in CV decreased. Otherwise, HSR led to poor economic performance. The Triangle of Central China (ID 4), Chengdu-Chongqing (ID 5), Central Guizhou (ID 14), west side of the Straits (ID 8), and Guanzhong Plain (ID 11) experienced a significant increasing trend with respect to POG, while the POG decreased in Harbin-Changchun (ID 9) and Mid-Southern Liaoning (ID 10), located in the northeast region. The coverage of HSR for urban agglomerations located in the northeast region is high, but the share of the economy continued to shrink. Therefore, with the development of transportation in the northeast region, integrating with other macro-policies to attract and retain productive economic activities is important. The urban agglomerations named Chengdu-Chongqing (ID 5), Central Guizhou (ID 14), and Guanzhong Plain (ID 11) are located in the western region. The share of the economy in the western region is growing. The economic disparity measured by the CV of PGDP in Chengdu-Chongqing (ID 5) and Central Guizhou (ID 14) decreased, while the disparity increased in Guanzhong Plain. The latest comprehensive transportation plan, published in 2021, took pride in the transportation development in Chengdu-Chongqing urban agglomerations, which has been regarded as the first economic growth pole in the western region.

Table 7.

Descriptive Statistics of Accessibility and Economic Indicators in Urban Agglomerations

	2007					2012					2018
ID	POG (%)	PGDP	ATT	ATI	DA4	POG (%)	PGDP	ATT	ATI	DA4	POG (%)	PGDP	ATT	ATI	DA4
1	9.6	25,436.3 (0.58)	1,032.4	827.3	21.4	9.2	45,719.9 (0.53)	754.5	355.0	45.7	8.7	61,804.2 (0.56)	549.6	339.6	62.6
2	19.5	35,861.0 (0.47)	1,113.3	719.6	15.9	18.1	64,728 (0.40)	792.6	342.0	37.2	19.3	101,865.7 (0.37)	549.6	289.4	74
3	9.8	36,532.3 (0.65)	1,332.9	633.1	10.2	8.7	57,572.7 (0.59)	863.1	293.5	17.6	9.3	87,590.3 (0.59)	612.2	253.6	45.9
4	7.0	16,047.5 (0.46)	991.8	651.3	17.0	8.5	39,246.8 (0.46)	755.7	354.7	37.3	9.2	63,100.8 (0.43)	527.9	276.5	76.8
5	5.0	12,699.4 (0.42)	1,275.5	1,401.6	8.3	5.6	29,037.7 (0.31)	1,049.9	872.9	9.6	6.2	47,085.4 (0.33)	671.7	546.9	22.2
6	7.7	15,271.4 (0.42)	928.6	721.6	26.6	7.4	30,660.2 (0.39)	755.6	376.8	47.7	7.5	45,886.3 (0.36)	514.5	320.9	75.8
7	6.0	45,353.9 (0.42)	1,161.2	933.5	12.4	5.3	78,264.5 (0.37)	869.7	471.2	27.1	5.3	114,309.8 (0.32)	633.5	421.3	37.9
8	5.3	20,558.2 (0.53)	1,312.1	772.4	6.4	5.2	40,721.6 (0.38)	990.2	445.5	12.1	5.8	68,319.9 (0.38)	649.3	309.3	42.2
9	3.7	22,198.6 (0.69)	1,547.7	1,338.1	15.2	3.8	50,297.9 (0.63)	1,058.8	675.9	20.2	2.8	53,124.8 (0.47)	830.6	623.6	28.0
10	3.9	31,677.6 (0.37)	1,339.2	1,183.6	17.2	4.0	64,749.8 (0.31)	909.2	548.0	26.2	2.4	58,144.8 (0.41)	709.6	536.7	31.0
11	1.8	11,568.0 (0.44)	1,102.6	1,126.6	8.9	2.0	27,316.2 (0.40)	856.9	592.4	13.7	2.1	40,730.0 (0.48)	623.8	507.3	35.6
12	2.0	14,230.5 (0.32)	1566.5	1,097.1	5.7	2.0	31,012.8 (0.28)	1,269.0	858.6	6.7	2.1	50,603.9 (0.26)	845.2	532.9	20.9
13	1.1	14,882.4 (0.38)	1,597.6	1,592.4	6.8	1.2	30,777.0 (0.37)	1,426.3	1,366.6	6.6	1.2	48,567.2 (0.36)	711.9	503.4	18.8
14	0.8	7,817.7 (0.68)	1,246.1	1,200.0	7.3	0.9	20,145.7 (0.43)	1,077.3	943.5	7.3	1.3	43,443.5 (0.39)	617.9	393.6	41.0
15	0.6	39,900.3 (0.74)	2,147.3	2,473.8	6.4	0.7	71,612.5 (0.49)	1,932.3	1,943.6	7.6	0.6	97,501.6 (0.29)	1,260.8	1,245.6	7.8
16	0.3	17,711.0 (0.47)	1,270.4	1,362.1	8.8	0.4	39,759.8 (0.41)	992.5	776.5	10.3	0.4	58,232.5 (0.38)	837.7	752.0	10.3
17	1.5	48,151.7 (0.38)	1,248.8	1,237.2	7.5	2.1	118,127.5 (0.35)	972.5	679.9	11.0	1.5	122,434.8 (0.28)	796.3	647.8	13.5
18	0.6	10,697.0 (0.53)	1,365.3	1,592.0	6.9	0.6	24,518.9 (0.47)	1,178.4	1,097.2	6.9	0.6	34,175.6 (0.53)	725.1	648.4	14.2
19	0.9	18,542.5 (0.57)	1,071.1	977.4	12.5	0.9	34,330.8 (0.36)	776.2	428.8	26.8	0.8	49,744.0 (0.45)	567.5	407.3	41.3

Note : POG = the proportion of the gross domestic product (GDP) of urban agglomerations to the total GDP of all research units; PGDP = gross domestic product per capita (yuan); ATT = average travel time to all prefectural level regions (min); ATI = average travel time to important cities (min); DA4 = daily accessible prefectural level regions (units).

Results

Associations of accessibility with regional economic productivity were estimated at national and four regional scales using different accessibility indicators. The regression results are provided in Table 8.

Table 8.

Regression Results of Three Accessibility Indicators

	Nation	East	Central	West	Northeast
Regression results of ATT
lnATT	−0.7964*** (−15.30)	−0.3233*** (−5.37)	−0.8252*** (−8.48)	−0.6081*** (−6.13)	−1.3665*** (−14.38)
lnPINV	0.2990*** (16.09)	0.3919*** (12.66)	0.3513*** (10.78)	0.3276*** (9.34)	0.1436*** (4.25)
lnPEOST	0.0910*** (6.79)	0.1413*** (6.89)	0.0238 (1.01)	0.1668*** (6.67)	0.0757** (2.37)
SPG	0.0098*** (7.46)	0.0080*** (4.41)	0.0113*** (5.00)	0.0086*** (3.67)	0.0133*** (5.23)
lnPNEP	0.0615** (2.05)	0.1241*** (3.58)	0.0521 (1.18)	0.0663 (1.00)	0.1653 (1.59)
AFF	0.0002 (1.08)	0.0002* (1.48)	0.0009* (1.68)	0.0003 (0.71)	−0.0003 (−0.34)
Constant	12.1334*** (26.49)	8.1095*** (12.27)	11.7214*** (14.48)	10.3109*** (11.42)	18.0623*** (19.94)
Hausman test	130.44***	6.45	34.61***	32.39***	26.74***
Estimated method	Fixed	Random	Fixed	Fixed	Fixed
No. of observations	864	261	240	252	111
R-squared	0.9412	0.9447	0.9632	0.9498	0.9447
Regression results of ATI
lnATI	−0.3604*** (−11.82)	−0.0562* (−2.12)	−0.3140*** (−6.72)	−0.4452*** (−6.93)	−0.9084*** (−11.58)
lnPINV	0.3517*** (18.80)	0.4402*** (12.63)	0.3967*** (12.00)	0.3017*** (8.56)	0.1558*** (3.94)
lnPEOST	0.1097*** (7.79)	0.1743*** (6.53)	0.0723*** (3.05)	0.1725*** (7.09)	0.0150 (0.38)
SPG	0.0005 (0.47)	0.0050** (2.01)	−0.0005 (−0.30)	0.0062*** (3.12)	0.0035 (1.38)
lnPNEP	0.0667** (2.09)	0.1364*** (2.85)	0.0671 (1.43)	0.0954 (1.48)	−0.0346 (−0.29)
AFF	0.0005** (2.20)	0.0005* (1.82)	0.0010* (1.69)	0.0002 (0.54)	0.0011 (1.26)
Constant	8.8694*** (27.23)	5.7841*** (9.49)	8.1610*** (15.79)	9.5340*** (13.54)	14.6535*** (17.30)
Hausman test	114.11***	14.31**	37.34***	34.68***	33.88***
Region effect	Fixed	Fixed	Fixed	Fixed	Fixed
No. of observations	864	261	240	252	111
R-squared	0.9334	0.9414	0.9582	0.9523	0.9248
Regression results of DA4
DA4	−0.0001 (−0.18)	0.0005 (0.62)	0.0010 (1.33)	0.0021 (1.29)	0.0245*** (3.52)
lnPINV	0.4734*** (26.84)	0.4547*** (14.56)	0.4883*** (14.29)	0.4660*** (15.94)	0.3336*** (5.94)
lnPEOST	0.1400*** (8.58)	0.1801*** (6.91)	0.0853*** (3.07)	0.1752*** (6.36)	0.1595*** (2.75)
SPG	−0.0041*** (−3.41)	0.0055** (2.11)	−0.0003 (−0.13)	−0.0003 (−0.15)	−0.0033 (−0.77)
lnPNEP	0.0888** (2.50)	0.1431*** (3.01)	0.0691 (1.31)	0.0881 (1.20)	−0.2386 (−1.27)
AFF	0.0008*** (3.47)	0.0005* (1.83)	0.0012* (1.81)	0.0008* (1.83)	0.0019 (1.31)
Constant	5.5136*** (30.81)	5.2400*** (15.92)	5.2591*** (16.98)	5.1500*** (15.36)	5.6801*** (8.93)
Hausman test	79.73***	16.65**	21.56***	32.47***	32.79***
Region effect	Fixed	Fixed	Fixed	Fixed	Fixed
No. of observations	864	261	240	252	111
R-squared	0.9171	0.9411	0.9466	0.9388	0.8110

Note : lnATT = natural log of average travel time to all prefectural level regions; lnATI = natural log of average travel time to important cities; DA4 = daily accessible prefectural level regions; lnPINV = natural log of the fixed-asset investment per capital; lnPEOST = natural log of expenditure for science and technology per capita; SPG = the share of production in the secondary industry; lnPNEP = natural log of the proportion of employed individuals to the total population; AFF = daily flight frequency. t-statistics are reported in parentheses below the estimated coefficient.

p < 0.1; **p < 0.05; ***p < 0.01.

First, we analyzed the regression results from the travel cost indicators, ATT and ATI. At the national level, the contributions from investments in ﬁxed assets per capita and expenditures for science and technology per capita were statistically significant. This confirmed that the fixed-asset investments and expenditures for science and technology had greatly contributed to regional economic productivity in China. In relation to the impact of improved accessibility through HSR development, ATT and ATI estimates were significantly negative. Therefore, accessibility improvements positively contributed to economic growth from 2007 to 2018 in China. The results showed that a 1% reduction in average travel time raised the GDP per capita by 0.796%, whereas a 1% decrease in ATI cities was associated with a 0.36% growth of GDP per capita. The assessment results at the four regional levels found that the fixed assets per capita also had a significant positive effect on regional economic productivity. The value of the coefficient was highest in the eastern region and lowest in the northeast region. This finding implies that the output efficiency of fixed assets was low in the northeast region. The reduction in travel time significantly contributed to each regional economic productivity. The influence coefficients of ATT were ranked from highest to lowest in the four regions as northeast, central, west, and east. The coefficients of ATI were, from highest to lowest, northeast, west, central, and east. The improved HSR accessibility had a larger effect on developing regions compared with the more developed east region. In particular, the reduction in travel time to important cities had a larger effect on the northeast and west regions. Accessibility improvement can contribute to alleviating regional disparity. In addition, flight frequency was found to be statistically significant in the eastern and central regions, but not in the western and northeast regions.

The DA4 indicator results showed that the only significant positive impact from this variable occurred for the northeast region; the results were not significant at the national level or at the other three regional levels. The railway accessibility indicator, measured by daily accessible units, was not suitable for regression models.

Conclusions

The China Railway Corporation is responsible for operating and managing the conventional railway and the HSR railway. The railway plays an important role in developing regional economies, and the economic differences are significant in China. As such, it was important to further analyze the impact of introducing HSR by using more detailed data, at different levels, and from different perspectives. Understanding how HSR projects affect individual regions differently, and what types of regions are more likely to show gains or losses, is very useful for formulating regional development policies.

First, we reviewed the planning history of the HSR network, combined with an analysis of the expansion process of the actual HSR network. This analysis highlights the potential implications for other countries with a similar HSR system or countries that are planning to develop such a system.

We then assessed the impact of adding HSR services on the accessibility of each prefectural level region in China. By examining four economic regions and 19 urban agglomerations, we were able to compare accessibility and economic performance. The spatial structure of China’s railway network is characterized by a “core-periphery” pattern, with Zhengzhou as the center. The CV for ATT and ATI increased year by year, indicating the accessibility disparity became larger with HSR development. These findings are consistent with Jiao’s et al. research ( 16 ). The changes in accessibility indicators indicated that the east and northeast regions improved the most between 2007 and 2012, and the west region significantly improved during the second study period. With respect to the urban agglomerations, the Triangle of Central China, Chengdu-Chongqing, and Central Guizhou urban agglomerations performed well with HSR development. In contrast, the city groups of Harbin–Changchun and Mid–Southern Liaoning located in the northeast region did not perform as well.

Finally, we estimated the economic impact of accessibility improvements. Accessibility improvements had a positive impact on regional economic productivity at both national and regional levels. The lowest impact coefficient of accessibility was in the eastern region. This finding implies that improved accessibility can contribute to alleviating regional disparity by bringing a larger effect to underdeveloped regions. This was confirmed by the panel regression analysis after controlling for other key factors, including investments, societal and technological development, economic structure, human capital, and air transport supply. These findings were consistent with previous studies about China’s HSR, including a study by Chen and Haynes ( 15 ). However, the basic research unit of Chen and Hayne’s research was the province level. The fitness of the regression model was poor, owing to a small observation set. Conducting an assessment at a more disaggregated regional scale, such as at prefectural level, was more valuable. Moreover, the findings were different from another Chinese case study conducted by Jin et al. ( 11 ). That study found that HSRs can aggravate economic disparity by bringing a signiﬁcant positive contribution to economic growth in large- to mega cities, but the contribution is insignificant in small- to medium cities. In their research, the measurement of HSR for each unit was the number of HSR lines, a quantity measurement of HSR. We used the accessibility indicator to measure the quality of HSR development for each unit. The estimated results of the two measurements are different. In most cases, the estimated results of the variable indicating the quantity of HSR, such as the density of HSR and the number of HSR stations, are often insignificant ( 15 ).

The ﬁndings of these assessments may provide guidance for future decision making with respect to transportation investment and sustainable economic development. Each urban agglomeration should take full advantage of their own local conditions and HSR network to promote economic growth and reduce intraregional economic disparity. The construction plan and the scales of the HSR network should be coordinated in line with regional economic development. In the Harbin–Changchun and Mid–Southern Liaoning urban agglomerations, the government should strive to attract enterprises and more productive industry, rather than relying on the HSR construction to promote economic growth. Railway accessibility improvement had less of an effect on developed regions than developing regions, whereas air transport had a larger effect on developed regions. This finding suggests that the western region is still in the stage of further increasing railway accessibility.

In closing, this study considered “onboard time” on trains, and did not include transfer times when calculating the shortest travel time. Future research could consider additional accessibility indicators, such as weighted average travel time, potential accessibility, and the frequency of trains. Despite this design constraint, this study provides a multilevel analysis in the field of accessibility and the economic effects associated with expanding the HSR network.

Footnotes

Author Contributions

The authors confirm contribution to the paper as follows: study conception and design: H. Kato, J. Fan; data collection: J. Fan, Z. Yang; analysis and interpretation of results: J. Fan, H. Kato, Y. Li; draft manuscript preparation: J. Fan, H. Kato. All authors reviewed the results and approved the final version of the manuscript.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the projects of the National Key R&D Program of China, Grant/Award Number: 2018YFB1601301

ORCID iDs

Jing Fan

Hironori Kato

References

Chen

Hong

Hsiao

Do China’s High-Speed-Rail Projects Promote Local Economy?—New Evidence From a Panel Data Approach. China Economic Review, Vol. 44, 2017, pp. 203–226.

Sun

Mansury

Y. S.

Economic Impact of High-Speed Rail on Household Income in China. Transportation Research Record: Journal of the Transportation Research Board, 2016. 2581: 71–78.

Qin

‘No County Left Behind?’ The Distributional Impact of High-Speed Rail Upgrades in China. Journal of Economic Geography, Vol. 17, No. 3, 2017, pp. 489–520.

Shao

Tian

Yang

High Speed Rail and Urban Service Industry Agglomeration: Evidence From China’s Yangtze River Delta Region. Journal of Transport Geography, Vol. 64, 2017, pp. 174–183.

Murakami

Cervero

High-Speed Rail and Economic Development: Business Agglomerations and Policy Implications. Routledge, England, 2017.

Sato

Sicular

Rising Inequality in China. Cambridge University Press, New York, NY, 2013.

Fan

Chan-Kang

Road Development, Economic Growth, and Poverty Reduction in China. Research Report 138. International Food Policy Research Institute, Washington, D.C., 2005.

Wetwitoo

Kato

High-Speed Rail and Regional Economic Productivity Through Agglomeration and Network Externality: A Case Study of Inter-Regional Transportation in Japan. Case Studies on Transport Policy, Vol. 5, No. 4, 2017, pp. 549–559.

Chen

Xue

Rose

A. Z.

Haynes

K. E.

The Impact of High-Speed Rail Investment on Economic and Environmental Change in China: A Dynamic CGE analysis. Transportation Research Part A: Policy and Practice, Vol. 92, 2016, pp. 232–245.

10.

Heuermann

D. F.

Schmieder

J. F.

The Effect of Infrastructure on Worker Mobility: Evidence From High-Speed Rail Expansion in Germany. Journal of Economic Geography, Vol. 19, No. 2, 2019, pp. 335–372.

11.

Jin

Lin

K.-C.

Shi

Lee

P. T.

K. X.

Impacts of High-Speed Railways on Economic Growth and Disparity in China. Transportation Research Part A: Policy and Practice, Vol. 138, 2020, pp. 158–171.

12.

Spiekermann

Wegener

Accessibility and Spatial Development in Europe. Scienze Regionali, Vol. 5, No. 2, 2006, pp. 15–46.

13.

Shaw

S.-L.

Fang

Tao

Impacts of High Speed Rail on Railroad Network Accessibility in China. Journal of Transport Geography, Vol. 40, 2014, pp. 112–122.

14.

Cao

Liu

X. C.

Wang

Accessibility Impacts of China's High-Speed Rail Network. Journal of Transport Geography, Vol. 28, 2013, pp. 12–21.

15.

Chen

Haynes

K. E.

Impact of High-Speed Rail on Regional Economic Disparity in China. Journal of Transport Geography, Vol. 65, 2017, pp. 80–91.

16.

Jiao

Wang

Jin

Dunford

Impacts on Accessibility of China’s Present and Future HSR Network. Journal of Transport Geography, Vol. 40, 2014, pp. 123–132.

17.

Martín

J. C.

Reggiani

Recent Methodological Developments to Measure Spatial Interaction: Synthetic Accessibility Indices Applied to High-Speed Train Investments. Transport Reviews, Vol. 27, No. 5, 2007, pp. 551–571.

18.

Hou

S.-M.

Transport Infrastructure Development and Changing Spatial Accessibility in the Greater Pearl River Delta, China, 1990–2020. Journal of Transport Geography, Vol. 19, No. 6, 2011, pp. 1350–1360.

19.

Pol

P. M.

The Economic Impact of the High-Speed Train on Urban Regions. Proc., 43rd European Regional Science Association ERSA Congress, Jyväskylä, Finland, 2003.

20.

Tomaney

Marques

Evidence, Policy, and the Politics of Regional Development: the Case of High-Speed Rail in the United Kingdom. Environment and Planning C: Government and Policy, Vol. 31, No. 3, 2013, pp. 414–427.

21.

Sasaki

Ohashi

Ando

High-Speed Rail Transit Impact on Regional Systems: Does the Shinkansen Contribute to Dispersion?

The Annals of Regional Science, Vol. 31, No. 1, 1997, pp. 77–98.

22.

Kim

Sultana

The Impacts of High-Speed Rail Extensions on Accessibility and Spatial Equity Changes in South Korea From 2004 to 2018. Journal of Transport Geography, Vol. 45, 2015, pp. 48–61.

23.

Saw

S.-H.

Wong

Regional Economic Development in China. Institute of Southeast Asian Studies, Singapore, 2009.

24.

Cao

Xue

Yan

A Study on the Urban Accessibility of National Trunk Highway System in China. Acta Geographica Sinica, Vol. 60, No. 6, 2005, pp. 25–32.

25.

Gutiérrez

Location, Economic Potential and Daily Accessibility: An Analysis of the Accessibility Impact of the High-Speed Line Madrid–Barcelona–French Border. Journal of Transport Geography, Vol. 9, No. 4, 2001, pp. 229–242.

26.

Liang

Zhou

Sun

Zeng

Effectiveness of High-Speed Railway on Regional Economic Growth for Less Developed Areas. Journal of Transport Geography, Vol. 82, 2020, p. 102621.