Intra-city access to inter-city transport nodes: The implications of high-speed-rail station locations for the urban development of Chinese cities

Abstract

Chinese

In the high-speed-rail (HSR) construction boom of China, although some cities have upgraded old train stations in inner cities to be compatible with HSR, more cities have built new HSR stations on undeveloped land in the urban periphery. This study investigates the impact of intra-city access to inter-city transport nodes and explores the implications of HSR station locations for the accessibility and residential property values in Chinese cities connected by bullet trains. We find that for the cities with HSR stations in suburbs, the gains in inter-city travel brought by HSR are largely offset by the prolonged intra-city travel time to reach the stations, thus limiting frequent usage of HSR for daily commuting. The inner-city HSR station in Hangzhou shows a positive impact on residential property value in the vicinity, while the suburban HSR station in Guangzhou has not been observed to raise the residential property values noticeably in the short term despite the government’s intention to stimulate development in surrounding areas. The research findings show the need for better connections of HSR stations with the city to magnify the accessibility provided by HSR and careful integrated planning to promote desirable urban development outcomes in station areas.

Keywords

accessibility China high speed rail property value station location

Research background

China has experienced an astonishing development of high speed rail (HSR) in recent years. In 2004, the Chinese State Council approved the ‘Mid-to-Long Term Railway Network Plan’, which sets the target at a total of 12,000 km of HSR lines by the year 2020.¹ This target was revised upwards to 16,000 km in 2008, and the implementation of the plan has been accelerated since then. As of 2014, China already has the largest HSR network in the world with over 11,800 km HSR lines.² It is projected that China’s HSR system will connect over 250 cities and carry 4 billion travellers annually by 2020 (Lou et al., 2011).

China’s ever-expanding HSR network is transforming the country both economically and socially. The rapid development of HSR has greatly increased the inter-city accessibility in a country with nearly 1.4 billion people geographically distributed in a land covering approximately 9.6 million km², facilitating exchanges and interactions among cities and regions (Zhu et al., 2016). In the meantime local governments also consider HSR an engine for boosting the urban development of municipalities. Many cities are planning to use new HSR stations as anchors to develop ‘HSR New Towns’ in the vicinity.³

The development effect of transport investment has been well investigated by a large number of studies on intra-urban transit stations that improve accessibility to job centres within a metropolitan area, such as subway stations, light rail stations and commuter rail stations. If accessibility to other cities’ markets also contributes significantly to the economic performances of firms and individuals, then areas in the vicinity of inter-city transport nodes should particularly benefit from regional integration. The intra-city trips to and from inter-city transport nodes within both the origin and destination cities could account for a significant portion of inter-city travellers’ total travel time. Similar to the case of intra-city transport stations, the potential reduction in transport cost could be capitalised into higher property values in areas close to inter-city transport nodes. However, research that examines the impact of access to inter-city transport nodes on an intra-city scale is relatively rare.

As a transport mode that mainly serves inter-city passenger travel, HSR needs to be integrated with local transport systems to provide door-to-door services to travellers. HSR stations serve as hubs that connect the inter- and intra-city transport systems, whose locations affect the intra-city access to HSR services. As in the cases of many European cities, a few Chinese cities chose to upgrade old train stations in inner cities to be compatible with the HSR system. However, more cities have built new HSR stations on undeveloped land in the urban periphery, which have weak connections to existing urban centres and internal transit systems of cities. China’s rapid HSR development provides a unique setting to investigate the intra-city access to inter-city transport nodes and understand the potential implications of different HSR station location strategies for the urban development of Chinese cities connected by bullet trains.

In this study, we evaluate the ‘net’ accessibility benefit brought by HSR to travellers from a door-to-door travel perspective taking into account the intra-city travel time variations as a result of different HSR station location strategies, and assess the associated development effect of HSR stations as reflected in the property value dynamics of surrounding residential neighbourhoods in the short term. The research findings could assist in the successful operation of HSR and in achieving desirable urban development outcomes in HSR cities by better informing decision makers of the potential effects. As multiple countries such as India, Thailand, Malaysia and Singapore are planning to build their own HSR lines, our study has policy implications not only for China but for other countries as well.

Related studies

Transport and urban development are closely related to each other in urban systems. Investments in transportation infrastructure have often led to the restructuring of urban land uses and land price patterns due to changes in accessibility. Conventional urban economics theory suggests that households make trade-offs between commuting cost and housing consumption in maximising their utility; therefore, residential property values increase with decreasing distance to job centres (Alonso, 1964). From an urban planning perspective, Webster (2010) argues that accessibility is a scarce resource that planners and urban designers seek to supply, and land value measures the benefit derived from accessibility. Many empirical studies in the USA have confirmed that proximity to traditional transit stations including subway stations, light rail stations and commuter rail stations has a significant positive effect on residential property values because of the reduced commuting cost (e.g. Diao, 2015; Diao and Ferreira, 2010; Gibbons and Machin, 2005; Hess and Almeida, 2007; McMillen and McDonald, 2004). Pan and Zhang (2008) find a similar pattern in China by analysing the accessibility impact of subway stations on housing prices in Shanghai. However, some researchers hold opposite views. For example, using California and Miami as case studies, Gatzlaff and Smith (1993) as well as Cervero and Landis (1997) find that development activities have not occurred around railway stations and land values of accessible locations are not higher than remote locations. Negative externalities such as noise and crime rate are among the reasons that could lead to the price decrease in the vicinities of stations (Bowes and Ihlanfeldt, 2001; Diao et al., 2015; Plano, 1993; Poister, 1996). One common characteristic of these studies is that they focus on intra-city transit modes and assess property value changes as a result of the commuting benefit brought by such modes.

As a significant transport innovation, HSR has many unique characteristics. Unlike traditional transit modes such as subway and light rail that facilitate better connectivity within an urban area and commuter rail that connects the middle to outer suburban areas with the urban core, HSR reduces the spatial barrier between cities. With an average speed range of 200–350 km/h, it is widely recognised that HSR is advantageous to other modes of inter-city travel such as automobiles and airlines for trips between 250 km and 900 km.⁴ Harman (2006) further defines three travel bands for HSR: commuting market (1 hour or less), primary market (1.5–2.5 hours) and longer-distance market (over 2.5 hours). Therefore, HSR could serve both short daily commuting trips and business and tourism trips at longer distances.

One concept relevant to HSR’s potential role in facilitating regional integration and promoting urban development is super-commuting – the phenomenon where employees work in one city and live in another city. Super-commuting has become the choice of many workers for multiple reasons such as lifestyle choices, high property values in big cities and pressure in the job market (Moss and Qing, 2012).⁵ Within certain travel time constraints, HSR could provide a viable commuting option for workers who need to travel between cities regularly. This trend is partly confirmed by Garmendia et al. (2012), who observe a significant increase in inter-city commuting in those HSR-connected cities in Spain, especially in intermediate cities. If HSR indeed supports such inter-city commuting, super-commuters may pay a premium for the accessibility benefit brought by HSR. In contrast, individuals who use HSR mostly for infrequent and long-distance business or tourism trips are less likely to pay. In the meantime, the accessibility benefit of HSR has to be assessed from a door-to-door travel perspective. The total travel time of HSR riders includes the inter-city travel time between a pair of HSR stations and the intra-city travel times to and from the HSR stations in the origin and destination cities, respectively. The location of HSR stations could influence travellers’ overall door-to-door travel time and experience by affecting the intra-city segments of travel. For example, using the Yangtze River Delta as a case study, Wang et al. (2013) conclude that travelling to and from HSR stations costs much more time for public transit riders than for travellers in private cars since most stations are located far away from city centres. According to a number of empirical studies from Western Europe and North America, commuting frequencies tend to drop beyond the critical 50-minute, one-way threshold travel time (Andersson et al., 2010). Therefore, entrenched intra-city travel time could rule out the possibility of super-commuting and constrain the accessibility benefit of HSR to business and tourism trips (Wang et al., 2013), thus reducing the attractiveness of HSR as a mobility tool and affecting the land use and property value effects of HSR stations in the immediate vicinity.

Despite the significant role of HSR in urban development and real estate markets, empirical studies on the impacts of access to long-distance rail services including HSR on residential property values are still limited and have produced mixed results. Some studies suggest a positive correlation between access to HSR stations and property values. For example, Haynes (1997) finds that the opening of HSR line between Paris and Le Mans led to a significant increase in land values in Le Mans. Meanwhile, he also observes the great demand growth for commercial spaces in areas around the Kassel station in Germany. Rietveld et al. (2001) find that residents in Vendome, France take into account the presence of HSR station in their residential location choices, as the introduction of HSR greatly reduced the travel time to Paris. Zheng and Khan (2013) investigate the impact of HSR-induced accessibility change on the housing price indices in 262 Chinese cities and find that cities with larger accessibility gains enjoy higher housing price appreciations. Their study provides useful insights into the property value effect of HSR, but as a city-level analysis, it cannot reveal the impact of HSR stations on the spatial patterns of property values within a city. Yin et al. (2015) provide a comprehensive summary on the international experiences of the direct and indirect development effects of HSR at the regional, urban and station-area levels and discuss the general implications of HSR on the urban development in China. They also point out that HSR stations could lead to an increase in property values in the station areas. On the other hand, another group of researchers argue that HSR has no significant impact on the real estate market in the station areas. Andersson et al. (2010) study Taiwan’s new HSR line and find that HSR accessibility has at most a minor effect on housing prices. The authors argue that high ticket prices and entrenched residential location patterns prevent otherwise feasible daily commuting opportunities between Tainan and other cities. Ahlfeldt’s (2012) study on Berlin’s Hauptbahnhoff station also suggests that HSR has no significant impact on the real estate market in the station neighbourhoods.

In summary, previous studies provide thorough analyses of the impact of intra-city transit stations on urban development and land valuation as a result of accessibility improvement to job centres in a metropolitan area. However, a detailed study on the development effect of inter-city HSR within a city and the potential impact of different HSR station location strategies is still lacking. China’s unique setting offers us a rare opportunity to systematically examine the intra-city access to inter-city transport nodes and the location models of HSR stations in a country with a rapidly expanding HSR network.

Location models of HSR stations in China

The majority of existing HSR stations in Europe and Asia are built either by upgrading old stations for conventional trains to be compatible with HSR or through new construction, thus leading to two location models: upgraded HSR stations are often located in developed urban areas, while newly built stations tend to be placed in the urban periphery. Upgrading existing stations saves the cost of land acquisition but greatly constricts the alignment of HSR lines and inconveniences the operation of the rail system during construction. In contrast, building new stations would have little interference on the current rail system during construction, but the choice of station location is worth careful consideration to balance the cost of development and the convenience of travellers. Unlike in the European countries that have many upgraded downtown HSR stations, the periphery model prevails in China.

In this section, we summarise the location patterns of HSR stations in Chinese cities resulting from the two different station construction models and evaluate the subsequent influences on the intra-city segments of HSR trips. We focus on all the 93 HSR stations connected by the G-series (Gaosu) and C-series (Chengji) trains in Mainland China. Both the G-series and C-series trains provide HSR services with a peak speed of over 300 km/h, the highest in the rail system of China. Among the 93 stations, 15 are upgraded stations and 78 are newly built stations; 66 stations are located in prefecture-level cities, while the remaining 27 stations are in county-level cities.⁶ To describe the relative location of an HSR station within a city, we compute a hypothetical radius for each HSR city (including both prefecture- and county-level cities) and compare the distance between a station and the corresponding city centre⁷ to this radius. Assuming that the built-up area of a city i is a circle,⁸ the hypothetical radius of the city R_i can be computed as:

R_{i} = \sqrt{A_{i} / π}

(1)

where A_i is the area of the built-up area of city i. It should be noted that this simplification may introduce biases due to the fact that each city has its unique shape. Figure 1 plots the relationship between the distance to city centre and the hypothetical radius of city. The dark line in Figure 1 has a slope of 1. HSR stations below this line are within the hypothetical boundaries of corresponding cities. It can be observed that all upgraded stations are below this line while the majority of new stations are above the line with several new stations located far beyond the hypothetical radius of city.

Figure 1.

Distance from HSR station to urban centre and radius of city.

Figure 2(a) and (b) plots histograms of two location measures for new stations and upgraded stations, respectively: the straight-line distance to city centre and the ratio of the distance to city centre to the hypothetical radius of city, which represents the relative location of an HSR station in a city. The figures reveal that new stations are generally further away from corresponding city centres in terms of both absolute and relative distances compared with upgraded ones. The weak linkage between suburban HSR stations to the city is further exacerbated by the poor public transport services connecting them. Public transport operators have low incentives to provide services to suburban HSR stations because of the low ridership as a result of the often low frequency of HSR services, the remote location from the urban centre, and the immature neighbourhoods and poor public services in the vicinity. We estimate the approximate travel time between each HSR station and corresponding city’s urban centre by public transit and private cars, respectively, using Google Maps. Google Maps provides functions of estimating travel time between two locations by transport modes based on the road network, normal travel speed and the schedule of public transport services. Figure 2(c) and (d) shows the histogram of the travel time from HSR stations to corresponding city centres by private cars and public transport, respectively. It can be observed that new stations have a significantly longer travel time than upgraded stations, especially for transit riders.

Figure 2.

Linkage between HSR stations and city centres.

Table 1 presents statistics that summarise the spatial relationship between HSR stations and city centres. On average, the newly built stations are located 11.23 km away from corresponding city centres, compared with 4.12 km for upgraded stations as measured by straight-line distances. The distances from newly built stations to city centres are approximately 2.15 times the hypothetical radius of city compared with 0.42 times for upgraded ones. In the meantime, it takes transit riders approximately 1.5 hours from a newly built station to the city centre and over 20 minutes for private cars under free-flow traffic conditions, which are significantly longer than the travel time of upgraded stations. It is demonstrated that newly built HSR stations have a substantially weaker connection with city centres than upgraded ones owing to their distant locations and poor public transport provisions.

Table 1.

Spatial relationship between HSR stations and city centres.

	Upgraded stations			New stations
Variables	Mean	Median	Obs.	Mean	Median	Obs.
Distance to city centre (km)	4.12	3.60	15	11.23	8.52	78
Ratio of distance to city centre and city radius	0.42	0.38	13	2.15	1.55	72
Travel time to city centre by public transport (min)	42.23	41.00	13	105.80	90.00	59
Travel time to city centre by private cars (min)	14.00	16.00	15	23.82	21.00	78

Multiple factors could have contributed to the adoption of the periphery model in Chinese cities. On the one hand, the remote location of new HSR stations could reduce the disputes between the government and local citizens on land acquisition, the potential delay due to over-complicated environment assessments, and the negative externalities during the construction period, thus facilitating the construction process. On the other hand, an HSR station is often seen as a new engine to boost urban development and economic growth. Locating anchor institutions such as universities and industrial parks at key suburban locations to induce growth in the surrounding areas is a widespread urban development strategy in China (Yang and Gakenheimer, 2007). The HSR new town is another example. The relatively lower acquisition cost and higher appreciation potential of suburban land mean higher revenues for the local government to reap once the station area evolves into a new town.

The accessibility impact of HSR station locations

In this section, we investigate the impact of HSR station locations on accessibility by comparing the inter-city travel time (as measured by weighted average travel time) savings as a result of HSR services and the intra-city travel time between HSR stations and city centres in prefecture-level Chinese cities.

Accessibility measure

We use weighted average travel time (WATT) to measure accessibility in this study, which is defined as the weighted average travel time between each city and all other cities (Gutiérrez, 2001). The WATT of a city can be calculated with the following formula using GDP of the city as weight:

A_{i} = \frac{\sum_{j}^{m} (t_{ij} M_{j})}{\sum_{j}^{m} (M_{j})}

(2)

where A_i is the WATT of city i, t_ij is the travel time between cities i and j, and M_j is the GDP of city j.

To compute the WATT, we first derive a station-level rail travel time origin-destination (OD) matrix. A graph model is created based on the national train schedule in which each node represents a train station and each edge is the actual travel time of a rail service between a station pair. A shortest-path algorithm is applied to the graph model to create a station-level travel time OD matrix $T_{station}$ , defined as:

T_{station} = {[t_{gh}]}_{n * n} (g = 1, 2, \dots, n; h = 1, 2, \dots, n)

(3)

Note that each element in the OD matrix t_gh is computed by following the shortest path through multiple connections in the railway network between an origin station g and a destination station h with a transfer penalty of 2 hours. t_gh = 0 when g = h. n is the total number of stations.

Based on the station-level OD matrix, a prefecture-level OD matrix $T_{prefecture}$ is generated. $T_{prefecture}$ is a subset of $T_{station}$ . Each element in $T_{prefecture}$ (t_ij) is the minimum travel time between a pair of railway stations that are located in the origin prefecture i and destination prefecture j, respectively. m is the total number of prefectures.

\begin{matrix} T_{prefecture} = {[t_{ij}]}_{m * m} \\ (i = 1, 2, \dots, m; j = 1, 2, \dots, m) \end{matrix}

(4)

t_{ij} = Min (t_{gh}) (\forall g ϵ i, \forall h ϵ j)

(5)

Inter-city travel time versus intra-city travel time

Previous accessibility analyses on HSR mainly focus on the inter-city travel time without considering the intra-city travel time to reach HSR stations. This simplification is reasonable for cities with HSR stations located in a central area. However, for cities with HSR stations far away from the urban centre, the travel time to and from the station could be a significant portion of travellers’ total door-to-door travel time and should not be neglected.

In order to understand the ‘net’ accessibility benefit brought by HSR, we construct two graph models based on national railway schedules before and after the introduction of HSR (in years 2009 and 2013, respectively) using the method described in section ‘Accessibility measure’, compute the savings in inter-city travel time for each of the 55 prefecture-level cities with HSR services, and compare it with the intra-city travel time between the major HSR station and the city centre. Among the 55 cities, eight have upgraded HSR stations while the other 47 have newly built stations.

Table 2 summarises statistics on intra-city travel times and inter-city travel time savings for the two groups of cities. We find that: (1) for travellers in cities with newly built HSR stations, the average travel time by public transport to reach the station is approximately 1.5 hours, almost twice the average travel time in cities with upgraded stations; and (2) in cities with newly built suburban stations, the gains in inter-city travel due to HSR are largely offset by the prolonged intra-city travel time to and from HSR stations, especially for transit riders. On average, the intra-city travel time by public transport is equivalent to 52.6% of savings in WATT for cities with new HSR stations compared with 23.9% for cities with upgraded HSR stations.

Table 2.

Intra-city travel time versus inter-city travel time savings in prefecture-level cities.

	Cities with upgraded HSR stations			Cities with new HSR stations
Variables	Mean	Median	Obs.	Mean	Median	Obs.
Savings in WATT (min)	198.60	211.80	8	195.80	175.90	47
Intra-city travel time (by car)	16.25	17.00	8	23.66	19.00	47
Intra-city travel time (by public transport)	45.50	45.38	8	93.38	84.50	42
Intra-city travel time (by car) as a percentage of WATT saving^a	8.43%	8.52%	8	13.34%	11.32%	47
Intra-city travel time (by public transport) as a percentage of WATT saving^a	23.88%	28.28%	8	52.57%	48.71%	42

Note: ^aThe percentage is first computed for individual cities. We then report the mean and median of the city-level percentages in Table 2.

With the prolonged travel time to urban centres, newly built suburban HSR stations tend to function like airports that mainly benefit infrequent business or tourism travels with a limited role in facilitating super-commuting. Several researchers have investigated the usage of certain HSR lines in China. In an on-board and station survey for passengers on the Shanghai–Nanjing HSR line, Tse et al. (2012) find that travelling for business, tourism and social purposes are the top purposes for using the HSR, while work accounts for only 5% of the journeys. Most passengers do not use the HSR on a regular basis, with only 8% of the passengers utilising it 2–3 days a week and 2% of the passengers using it 4–5 days a week. Wang et al. (2013) also report similar findings in their HSR passenger survey. These survey results are consistent with our accessibility analyses that take into account both the inter- and intra-city segments of HSR journeys. For super-commuters, living and working in locations close to HSR stations at both ends of commuting trips could mitigate the negative effect of intra-city travel, but new HSR stations tend to be built in suburban areas with a long distance from the urban centre and in the vicinity of immature neighbourhoods, thus limiting the feasibility of super-commuting.

The property value impact of HSR station location

In this section, we assess the impact of HSR stations on residential property values in two HSR cities with different station location strategies, Guangzhou and Hangzhou. Guangzhou and Hangzhou play key roles in the economic development of the Pearl River Delta and the Yangtze River Delta, respectively. The expansion of the HSR network across the country has brought significant impacts to the two cities and made them regional hub cities in the HSR network. To embrace the rapid development of HSR, Guangzhou built the Guangzhou South Railway Station, on the urban periphery approximately 17 km (straight-line distance) south of central Guangzhou. The municipal government is trying to use this newly built HSR station to enhance the development in the south Guangzhou area. According to the ‘Regulatory Plan for the Guangzhou South Station Area’,⁹ the planning area of this HSR new town is approximately 36 km² with a planned total population of 0.37 million. One component of the strategy of the Guangzhou government is to attract residents or investments from the neighbouring city, Foshan. The two cities are geographically close to each other, and share a common culture and history. The Guangzhou South Railway Station is expected to play a key role in facilitating the economic integration of the two cities. However, whether the new HSR station can indeed stimulate development in the vicinity as planned remains unclear. By comparison, Hangzhou’s HSR services are mostly concentrated in the Hangzhou East Railway Station, which is approximately 6 km east of the city centre and was upgraded from the old railway station in the urban area. Considerable redevelopment and new development activities are still expected to happen in this area. The locations of the two cities in the HSR network as well as the locations of the two HSR stations in the corresponding city are shown in Figure 3.

Figure 3.

Locations of the two cities and stations.

The different location choices of HSR stations in Guangzhou and Hangzhou offer us an opportunity to explore the role of HSR station locations in facilitating urban development as reflected by the dynamics in residential property values in the vicinity. Multiple factors could affect the transaction price of properties in the station areas. Conceivably, living adjacent to the HSR stations has the benefits of improved connectivity to other cities, which is particularly desirable for frequent HSR riders such as super-commuters. The newly developed or gentrified commercial areas around the stations may also be a drawing factor. However, as a major attraction point of intra-city traffic, the congestion, noises and emissions resulting from a large amount of car trips around the HSR stations would certainly discourage some people from living nearby. More crucially, people usually consider the distances from housing locations to other important destinations such as workplaces, schools and shopping centres when choosing the location of their future homes. Therefore, the spatial relationships between HSR stations and other activity centres also play a key role in attracting households to live in the station areas. Between the two HSR stations of interest, Hangzhou East Railway Station has better accessibility to such amenities as a result of its central location in the city, while Guangzhou South Railway Station has disadvantages in this respect as a newly built station on the urban periphery. Property value is an important indicator for assessing the extent to which the HSR system can help the development of local areas and for evaluating whether the enormous investment can be paid off.

Data sets

In this study, we collected resale housing transaction information from online sources. In the case of Hangzhou, the resale transaction data are at the housing project level, which are readily available from Soufun.com, a leading real estate advertising platform. The data include the average transaction price and housing unit size of major residential projects in Hangzhou from 2007 to 2013 and some project-specific characteristics such as green area ratio. Green area ratio is the percentage of neighbourhood lot area that is occupied by plants including trees, gardens, lawns and green roofs, etc. The resale transaction data for Guangzhou are from mytophome,¹⁰ which contain transaction information for individual housing units from 2005 to 2013. The data source does not provide information on neighbourhood environment, although this can be complemented from other sources because the housing project names are known.

Model specification

Hedonic price models have been used extensively in estimating how property values are influenced by observable property characteristics and neighbourhood attributes. In this study, we calibrate hedonic price models to estimate the effect of HSR stations and the resultant changes in inter-city accessibility and intra-city (re)development on residential property values, which can be formulated in the form of:

\begin{matrix} \ln (P_{ijt}) = & κ + α_{1} D_{i} + α_{2} D_{i}^{2} + α_{3} D_{i}^{3} + α_{4} D_{i} \\ \times Y_{t} + β N_{jt} + φ H_{it} + γ Y_{t} + ε_{ijt} \end{matrix}

(6)

where $P_{ijt}$ is the transaction price per square metre of property i of neighbourhood j in year t. $H_{it}$ is a vector of features specific to property i, which is assumed to be constant over time, except for the age of units. $N_{jt}$ is a vector of neighbourhood characteristics, and $D_{i}$ is the distance from property i to HSR station. To capture the nonlinear relationship between distance to HSR station and housing price, we adopt a spline function of $D_{i}$ in the specification. $Y_{t}$ is a set of dummy variables indicating the year of transaction to capture the temporal change in the housing market. In addition, $D_{i} \times Y_{t}$ is an interaction term between distance to HSR station and transaction year dummy variables in order to capture the yearly change of the effects of HSR station on property price. $ε$ is the error term. For the Hangzhou model, we use project-level characteristics as a proxy for the attributes of individual housing units. For example, the age of the projects is used as the age of units. $P_{ijt}$ is the average transaction price per square metre of all transactions in project j and year t. The statistical modelling takes the form of independently pooled panel analysis.

Model results

Table 3 shows the result of the hedonic price model for the city of Hangzhou. As expected, building age has a significant negative effect and unit area has a significant positive effect on transaction prices. The bigger the housing unit, the higher the average unit price per square metre. The green area ratio of a housing project, an indicator of neighbourhood environment, is also positively related with housing prices. Distance to the nearest metro station has a negative and significant coefficient, indicating that households would pay premiums for better accessibility to a metro station. Distance to West Lake, the main attraction of the city of Hangzhou, has a significant negative effect on housing prices, while the effect of distance to the city centre is positive but insignificant. This suggests that Hangzhou citizens are willing to live closer to West Lake than to the city centre.

Table 3.

Hedonic price model for Hangzhou.

Hangzhou	Estimate	Std err.	t value
Intercept	9.1240	0.0538	169.4750 ***
Property attributes $(H_{it})$ :
Building age (year)	−0.0720	0.0073	−9.8140***
Ln(unit area) (ln m²)	0.1358	0.0081	16.7430 ***
Neighbourhood attributes ( $N_{jt}$ ):
Green area ratio	0.0017	0.0002	8.7360 ***
Distance to CBD (km)	0.0080	0.0042	1.9210
Distance to the West Lake (km)	−0.0649	0.0032	−20.4420 ***
Distance to the nearest metro station (km)	−0.0192	0.0021	−9.2970 ***
Distance to HSR station ( $D_{i}$ ):
Distance to East Railway Station (km)	0.0097	0.0083	1.1690
Distance to East Railway Station² (km²)	−0.0024	0.0011	−2.1190 *
Distance to East Railway Station³ (km³)	0.0001	0.0000	3.1710 **
Transaction year dummies ( $Y_{t}$ ):
Year 2007 (reference)
Year 2008	0.1879	0.0253	7.4170 ***
Year 2009	0.3246	0.0244	13.3000 ***
Year 2010	0.6778	0.0233	29.1120 ***
Year 2011	0.8445	0.0240	35.1610 ***
Year 2012	0.7962	0.0254	31.3520 ***
Year 2013	0.8267	0.0278	29.7850 ***
Interaction terms ( $D_{i} \times Y_{t}$ )
Distance to East Railway Station * Year 2008	−0.0051	0.0042	−1.2120
Distance to East Railway Station * Year 2009	0.0002	0.0040	0.0570
Distance to East Railway Station * Year 2010	−0.0021	0.0037	−0.5590
Distance to East Railway Station * Year 2011	−0.0080	0.0037	−2.1360 *
Distance to East Railway Station * Year 2012	−0.0112	0.0039	−2.8540 **
Distance to East Railway Station * Year 2013	−0.0113	0.0042	−2.6890 **
Number of observations: 5764 Multiple R-squared: 0.7012, Adjusted R-squared: 0.7002
F-statistic: 641.8 on 21 and 5742 DF, p-value: < 2.2e−16

Notes: Significance codes: *** 0.001, ** 0.01, * 0.05, . 0.1.

In the case of Guangzhou, it has two HSR stations. The Guangzhou South Railway Station is the major one serving trains running on all the HSR lines connecting to Guangzhou, while the Guangzhou North Railway Station is only an intermediate station on the Wuhan–Guangzhou HSR line.¹¹ Because the Guangzhou North Railway Station only serves a few trains and has limited impact on the inter-city transportation of Guangzhou, we only consider the impact of the Guangzhou South Railway Station in the model. Table 4 presents the estimation results of the hedonic price model for Guangzhou. The effect of distance to the city centre is significantly negative, suggesting that the closer to the city centre, the higher the transaction price.

Table 4.

Hedonic price model for Guangzhou.

Guangzhou	Estimate	Std err.	t value
Intercept	6.3110	0.1075	58.6850***
Property attributes $(H_{it})$ :
Ln(area) (ln m²)	0.0878	0.0039	22.2980***
Neighbourhood attributes ( $N_{jt}$ ):
Distance to CBD (km)	−0.0064	0.0004	−15.6350***
Distance to HSR station ( $D_{i}$ ):
Distance to South Station (km)	0.2409	0.0080	29.9530***
Distance to South Station² (km²)	−0.0113	0.0002	−45.5270***
Distance to South Station³ (km³)	0.0002	0.0000	41.8960***
Transaction year dummies ( $Y_{t}$ ):
Year 2007	0.7925	0.1021	7.7630***
Year 2008	0.8442	0.1026	8.2260***
Year 2009	1.0340	0.1022	10.1140***
Year 2010	1.3010	0.1024	12.7090***
Year 2011	1.5210	0.1028	14.8000***
Year 2012	1.6090	0.1030	15.6260***
Year 2013	1.7890	0.1043	17.1580***
Interaction terms ( $D_{i} \times Y_{t}$ )
Distance to South Station * Year 2006	−0.0084	0.0065	−1.2940
Distance to South Station * Year 2007	−0.0094	0.0054	−1.7400
Distance to South Station * Year 2008	−0.0102	0.0064	−1.5930
Distance to South Station * Year 2009	−0.0120	0.0064	−1.8790
Distance to South Station * Year 2010	−0.0112	0.0064	−1.7510
Distance to South Station * Year 2011	−0.0177	0.0064	−2.7630**
Distance to South Station * Year 2012	−0.0106	0.0064	−1.6610
Distance to South Station * Year 2013	−0.0135	0.0065	−2.070*
Number of observations: 60,471 Multiple R-squared: 0.4941, Adjusted R-squared: 0.4939
F-statistic: 2187 on 27 and 60471DF, p-value: < 2.2e−16

Notes: significance codes: *** 0.001, ** 0.01, * 0.05, . 0.1.

The effect of distance to HSR stations on property values

Figures 4 plots the estimated effects of the HSR station on residential property values of surrounding areas based on the calibration results of the two hedonic price models for Hangzhou and Guangzhou, respectively. X axis shows the distance to HSR station and Y axis plots the estimated percentage difference in the unit price per square metre of housing units.

Figure 4.

Housing unit price difference in percentage by distance to the HSR station.

The Hangzhou East Railway Station is located in the urban area and was upgraded from the existing station. As the construction of the HSR network and the station proceeded, the impacts of the HSR station on housing prices also changed over the years. Before 2011 when the HSR line between Hangzhou and Shanghai was completed, immediate adjacency to the station has a negative effect on housing prices as shown in the housing price gradient because of the nuisance of noise and congestion brought by the construction. The housing price increases with the distance and reaches a peak in places approximately 2.5 km away from the station. Since 2011, with the renovation of the station close to finish and the gentrification of the nearby urban environment, the station has increasingly become an attraction factor as opposed to a repulsion factor for home buyers. The average unit price of housing projects decreases with distance to the station. Super-commuting could be a potential factor that contributes to the change in land gradient, as HSR could shorten the travel time between Hangzhou and Shanghai to approximately 50 minutes. Although Hangzhou is an important city in the Yangtze Delta Region, its role is not comparable with Shanghai, which provides better career opportunities and services. It is conceivable that a number of people living in Hangzhou are super-commuters or frequent visitors to Shanghai. Therefore, the neighbourhoods around the HSR station are particularly attractive to this group of people.

According to the estimation results, being closer to the Guangzhou South Railway Station tends to significantly lower transaction price. Similar to Hangzhou, the effect of the distance to the HSR station on housing prices has changed over the years. It is noteworthy that the gradient lines for the years after 2007 are generally flatter than the line for the years 2007, which might be explained by the growing expectation for the HSR service in the minds of both home buyers and speculators. When the HSR line between Guangzhou and Wuhan came into operation in 2009 and the Guangzhou South Railway Station was put in use in 2010, it temporarily stimulated the property values in surrounding areas as shown by the flatter black lines in Figure 4. However, because the Guangzhou South Railway Station is a newly built station at a suburban location far away from the city centre, its stimulation on property values in surrounding areas has been limited since 2007. The slow development of the surrounding areas might be attributed to the poor accessibility to other parts of the city and the lack of necessary services, facilities and jobs locally, which make the station areas unsuitable for serving as either origin or destination of super-commuting.

In a nutshell, the location of HSR stations in cities seems to be a factor that could affect the property value effects of HSR. The housing price appreciation around new suburban HSR stations is limited in the near term because of the weak connections with existing urban centres and the poor provisions of amenities and jobs. The upgraded HSR stations, in contrast, may help promote the values of nearby residential neighbourhoods. This could be the case especially for those HSR stations in medium and small cities.

Conclusions and policy implications

With the ongoing HSR construction boom in China, a growing number of Chinese cities have experienced an improvement in inter-city connections brought by bullet trains. In this study, we investigate the accessibility implications of HSR station locations in Chinese cities and the subsequent effects on urban development and property values in surrounding areas. This study contributes to a better understanding of the impact of intra-city access to inter-city transport nodes by shedding some new lights on critical issues such as whether market values the intra-city access to HSR stations and whether the location choice of HSR stations affects the land use and property value effects of HSR.

We find that during the expansion of the HSR network in China, although some cities upgrade existing railway stations in urban centres to be compatible with HSR, more cities tend to build new HSR stations in remote suburban sites. The weak linkage between distant suburban HSR stations and cities is further exacerbated by the poor public transport services connecting them.

HSR has greatly reduced people’s commuting time between Chinese cities. However, for cities with HSR stations in the suburbs, the gains in inter-city travel are largely offset by the prolonged intra-city travel from suburban HSR stations to city centres. The intra-city part of trips could be cumbersome for cross-city passengers. As a result, these stations and the HSR services are not necessarily beneficial for frequent short-distance users of HSR such as super-commuters, who particularly value the accessibility benefit, ease and cost savings associated with the whole door-to-door trip.

To understand the development effect of HSR stations at the intra-city scale, we employ hedonic price models to analyse the changes in residential property values in the vicinity of HSR stations in two Chinese cities with different station location strategies. The upgraded HSR station in the inner-city of Hangzhou appears to be a catalyst for the redevelopment of the station area and shows a positive impact on residential property values in the vicinity, while the newly built suburban HSR station in Guangzhou has not been observed to raise residential property values noticeably in the short term despite the government’s intention to stimulate development in surrounding areas. This could be partially attributed to its limited role in facilitating super-commuting.

Our findings have significant policy implications. As a new transport mode, HSR not only creates a shift in economic geography by shrinking inter-city travel time, but also has the potential to alter intra-city development dynamics of HSR cities. The location alternatives of HSR stations need to be carefully evaluated based on the specific situations of cities, such as population and economic growth, and long term urban development plan. Their corresponding challenges must be addressed.

Well-connected HSR stations in inner-city locations, especially when coupled with other investments, offer great potential for urban revitalisation. In spite of the many challenges in land availability and environmental externalities, inner-city HSR stations can bring economic vitality and redevelopment opportunities to ‘old’ urban areas that would be less attractive to developers without new catalysts.

Despite its popularity in China, the remote location of suburban stations has raised significant challenges for cities. One of the major goals for building the HSR network in China is to facilitate cross-city economic integration by improving market access, expanding the labour market, and enhancing spatial agglomeration. However, the weak linkage between suburban HSR stations and corresponding city centres hinders such integration. Better transit services (e.g. higher frequency, more routes and shorter travel time) are needed to connect suburban HSR stations to city centres more efficiently, thus magnifying the accessibility benefit brought by HSR.

Many Chinese cities plan to use new suburban HSR stations as an engine to develop HSR new towns. However, HSR in itself will not guarantee development around stations. The remoteness and the dearth of existing service centres and business activities could discourage home buyers to the station areas, thus ‘ghost towns’ could emerge. To attract home buyers to the new town, the development of appropriate job centres and the provision of local service centres, such as shopping, educational, recreational and public service facilities, should be coordinated and keep up with the pace of station and housing development in the station area. The new town needs to be largely self-sufficient so that its residents do not have to rely heavily on existing city centres to meet their activity needs.

In the meantime, the HSR-induced inter-city accessibility benefit is a unique advantage of HSR new towns, thus catering to the special needs of super-commuters and frequent business travellers in the planning, and development could be beneficial for the success of new towns. In small or medium cities that most likely serve as origin of super-commuting, planners should pay attention to the provision of necessary services and facilities in the station areas to meet the daily needs of super-commuters. In large or mega cities that serve as destinations of super-commuting, jobs especially those may attract worker from other cities need to be placed in areas close to HSR stations, or convenient intra-city connections between HSR stations and existing job centres should be provided.

In summary, inner-city HSR stations could stimulate urban redevelopment in the station areas, but also face challenges such as land availability and negative environmental externalities. The periphery model adopted by most Chinese cities has contributed to the rapid expansion of the HSR network in China but also brought many issues that need to be carefully addressed to magnify the accessibility benefit of HSR, promote economic integration and facilitate desirable urban development outcomes in the station areas. In the long term, with the change of city spatial structures and the improvement of transport connectivity between suburban stations and city centres, it is expected that the differences between these two types of stations could become smaller.

Several limitations of this study should be noted: (1) because of data limitations, the two hedonic price models for Guangzhou and Hangzhou have different explanatory variables. Furthermore, each model only includes several key variables, while some important factors that could influence property values are neglected, such as land use mix and accessibility to various amenities. Better data could improve the quality of our models. (2) We use two case study cities, Guangzhou and Hangzhou, to investigate the development implications of two HSR station location models. Extending the study to more HSR cities could help generalise our findings.

Footnotes

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Notes

References

Ahlfeldt

(2012) Rail mega-projects in the realm of inter-and intra-city accessibility: Evidence and outlooks for Berlin. Built Environment38(1): 71–88.

Alonso

(1964) Location and Land Use: Toward a General Theory of Land Rents. Cambridge, MA: Harvard University Press.

Andersson

Shyr

(2010) Does high-speed rail accessibility influence residential property prices? Hedonic estimates from southern Taiwan. Journal of Transport Geography18(1): 166–174.

Bowes

Ihlanfeldt

(2001) Identifying the impacts of rail transit stations on residential property values. Journal of Urban Economics50(1): 1–25.

Cervero

Landis

(1997) Twenty years of the Bay Area Rapid Transit System: Land use and development impacts. Transportation Research Part A: Policy and Practice31(4): 309–333.

Diao

(2015) Selectivity, spatial autocorrelation and the valuation of transit accessibility. Urban Studies52(1): 159–177.

Diao

Ferreira

(2010) Residential property values and the built environment: Empirical study in the Boston, Massachusetts, metropolitan area. Transportation Research Record: Journal of the Transportation Research Board2174: 138–147.

Diao

Qin

Sing

(2015) Negative externalities of rail noise and housing values: Evidence from the cessation of railway operations in Singapore. Real Estate Economics. DOI: 10.1111/1540-6229.12123.

Garmendia

Ribalaygua

Ureña

(2012) High speed rail: Implication for cities. Cities29: S26–S31.

10.

Gatzlaff

Smith

(1993) The impact of the Miami Metrorail on the value of residences near station locations. Land Economics69: 54–66.

11.

Gibbons

Machin

(2005) Valuing rail access using transport innovations. Journal of Urban Economics57(1): 148–169.

12.

Gutiérrez

(2001) Location, economic potential and daily accessibility: An analysis of the accessibility impact of the high-speed line Madrid–Barcelona–French border. Journal of Transport Geography9(4): 229–242.

13.

Harman

(2006) High Speed Trains and the Development and Regeneration of Cities. London: Greengauge21.

14.

Haynes

(1997) Labor markets and regional transportation improvements: The case of high-speed trains: An introduction and review. The Annals of Regional Science31(1): 57–76.

15.

Hess

Almeida

(2007) Impact of proximity to light rail rapid transit on station-area property values in Buffalo, New York. Urban Studies 44(5/6): 1041–1068.

16.

Lou

Gui

Moh

. (2011) China High-speed Rail: On the Economic Fast Track. New York: Morgan Stanley Research Global.

17.

McMillen

McDonald

(2004) Reaction of house price to a new rapid transit line: Chicago’s Midway Line, 1983–1999. Real Estate Economics32: 463–486.

18.

Moss

Qing

(2012) The emergence of the ‘super-commuter’. Report for Rudin Centre for Transportation, Wagner School of Public Service, New York University. Available at: http://wagner.nyu.edu/files/rudincenter/supercommuter_report.pdf.

19.

Pan

Zhang

(2008) Rail transit impacts on land use: Evidence from Shanghai, China. Transportation Research Record: Journal of the Transportation Research Board2048(1): 16–25.

20.

Plano

(1993) Transit generated crime: Perception versus reality - A sociogeographic study of neighborhoods adjacent to section B of Baltimore Metro. Transportation Research Record 1402: 59–62.

21.

Poister

(1996) Transit related crime in suburban areas. Journal of Urban Affairs18: 163–175.

22.

Rietveld

Bruinsma

Van Delft

(2001) Economic impacts of high speed trains: Experiences in Japan and France, expectations in The Netherlands. Research Memorandum 2001-20, Vrije Universiteit, Amsterdam. Available at: degree.ubvu.vu.nl/repec/vua/wpaper/pdf/20010020.pdf.

23.

Tse

Wang

Zhao

(2012) High speed rail and commuting patterns in Yangtze River Delta: An initial observation. Proceedings of the 2012 IEEE/ASME Joint Rail Conference, Philadelphia, Pennsylvania, USA, 17–19 April, 2012. New York: The American Society of Mechanical Engineers. Available at: http://proceedings.asmedigitalcollection.asme.org/proceeding.aspx?articleid=1716336.

24.

Wang

(2013) Spatial impacts of high-speed railways in China: A total-travel-time approach. Environment and Planning A45(9): 2261–2280.

25.

Webster

(2010) Pricing accessibility: Urban morphology, design and missing markets. Progress in Planning73: 77–111.

26.

Yang

Gakenheimer

(2007) Assessing the transportation consequences of land use transformation in urban China. Habitat International 31(3–4): 345–353.

27.

Yin

Bertolini

Duan

(2015) The effect of the high-speed railway on urban development: International experience and potential implications for China. Progress in Planning98: 1–52.

28.

Zheng

Kahn

(2013) China’s bullet trains facilitate market integration and mitigate the cost of megacity growth. PNAS 110(14): E1248–E1253.

29.

Zhu

Diao

(2016) The evolution of accessibility surface of China in the high-speed-rail era. Environment and Planning 48(11): 2108–2111.