The Effects of Subway Construction on Housing Premium: A Micro-data Analysis in Chengdu s Housing Market

The Effects of Subway Construction on Housing Premium: A Micro-data Analysis in Chengdu s Housing Market Cong Sun Siqi Zheng Rikang Han Abstract As a sign of city development and prosperity, subway is not only an important public transportation for residents travelling, but also an engine to real estate market and business booming around the stations. Moreover, the positive externalities brought by the subway will also be capitalized into the nearby properties. In this paper, the micro-data of new commodity housing units in Chengdu was used to estimate the premium effect during subway construction (Line 1 and 2) from year 2006 to the first half of year 2010. The empirical results suggest that the spatial and temporal housing premium are significant, and that the average home price is 7%-14% higher within 1.5km around the subway station than outside the stations. Based on above estimation, the premium of residential land is more than 5 billion U.S dollars during the period, which can cover the subway construction cost (2.86 billion U.S dollars). The empirical results also provide some policy implication for urban subway construction and local public finance. Keywords Subway Premium Home price Cong Sun Siqi Zheng ( ) Institute of Real Estate Studies, Tsinghua University, Beijing, P.R. China e-mail: zhengsiqi@tsinghua.edu.cn Cong Sun e-mail: suncong05@gmail.com Rikang Han Department of Real Estate & Construction, The University of Hong Kong, Hong Kong, P. R. China e-mail: hrk11@hku.hk

2 Introduction Chengdu has started the construction of Urban High-speed Track Transportation Network since 2005. Rail transit can not only provide passengers with rapid, convenient, punctual and green service but also shunt road traffic and improve urban overall traffic situations, thus enhancing urban overall traffic accessibility and reducing carbon emission of urban traffic. In the long run, rail transit can greatly improve the transportation accessibility of suburban area and create a more proper city spatial structure, which could promote agglomeration of city central areas and provide guarantee for further development. Of course, rail transit has a most remarkable effect around subway stations. With the improvement of transportation accessibility around subway stations, more and more residents as well as enterprises will be attracted to gather around subway stations, as a result more investments of commercial and public facilities will also be attracted there, thus economic vitality of surrounding areas of subway stations will be greatly promoted. In a word, subway construction will bring great social and economic benefits, which is especially significant around subway stations. And these benefits will be capitalized into surrounding property, resulting in the premium of the surrounding real estate market. But there are quite huge costs in rail transit construction, taking Chengdu as an example, in which the total investment of the first-phase construction of subway line 1 and 2 reaches up to 18.3 billion, so it is a hot issue for scholars to research on how to calculate the effects on surrounding real estate market caused by subway construction and internalize the positive externality brought by subway construction. This paper constructs spatial database based on geographic information system (GIS) with Chengdu newly built commodity housing. It aims at the construction situation of Chengdu Urban High-speed Track Transportation Network, references domestic and abroad research methods about subway effect analysis and economic evaluation, verifies the positive effects on surrounding housing market caused by the major subway line (Line 1 and 2) and quantitatively calculates the premium of housing as well as residential areas. On the above basis some suggestions are put forward on how to effectively utilize land premium to promote urban development construction as well as the stable and healthy development of real estate market. Literature Review Subway can bring premium for surrounding land and real estate housing. The economic mechanism is that subway foundation enhances traffic accessibility by providing people with more convenient trip; at the same time, people gather around rail transit site, arousing many kinds of business opportunities and attracting more investments, thus surrounding infrastructure can be improved. Under the

3 construction of subway, premium creates as residents and real estate developers expect that subway will bring benefits to the neighborhood in the future and the future profits will be capitalized into the surrounding home prices. The spatial difference of the premium is called "Spatial Effect". Moreover as time goes by, the premium degree will be larger because of more accurate information of subway and its attraction to public facilities. The premium change at different time is called "Time Effect". On the study of spatial effect, the international consensus is that on the premise of controlling other variables, the home price and the rent level will have a rising trend with a closer distance to subway (Amstrong,1994[1] ; Cervero and Duncan, 2002[2]; Richard,1993[10]). As subway lines in our country is rising, domestic research mainly focuses on cities which have had subways or under construction, and the research also finds that home price around subway will get rising (He and Wang, 2004[5]; Zheng and Liu, 2005[11]; Liang et al, 2007[8]; Pan and Zhong, 2008[9]; Gu and Zheng, 2010[4]). There are also some scholars doing calculation on increment brought by subway effect, of whom Zheng and Liu (2005) have a calculation result showing that increment profit of real estate with 500m radius range in each site of each phase of Shenzhen subway is 0.27 billion U.S dollars. 20 sites of each phase amounting to 5.23 billion U.S dollars and 2.9 times of the first phase subway construction cost (1.79 billion U.S dollars)[11]. Gu and Zheng (2010) have a calculation on premium brought by subway[4], of which the conservative estimate result is that Beijing subway Line 13 brings the premium of 8.2 billion U.S dollars for surrounding housing and they are enough to make up for the construction cost. As for the study of time effect, there are big differences among different projects. Knaap (2001) finds that light rail investment information has positive function on surrounding land price[7]; but Gatzlaff et al (1993) finds that the information promulgation of Miami new subway system at most has slight effect only[3]; however, research of Henneberry (1998) shows that Sheffield light rail project made a reduction of surrounding home prices at first, but the price had a recovery trend[6]. Data and Variables Data The projects under construction (or being operated) were Chengdu subway Line 1 and 2 till October 2010. These two subway lines were both approved to construct by the State Council on Aug. 13, 2005, and classified into three phases to implement. The total investment of the first phase engineering of Chengdu

4 subway Line 1 was approximately 1.3 billion U.S dollars with Shengxian Lake at its north and Century City at its south, running across major urban zones. Its overall length was 18.5 kilometers, which was all underground line. And it went to operation on December 27, 2005 and opened to traffic in October 2010; the overall length of the first phase of subway Line 2 was 23 kilometers, and the whole line extended from northwest to southeast. Its sites were 20 with the total investment being 1.45 billion U.S dollars. Subway Line 2 started operation on December 29, 2007 and was expected to be on trial operation in the end of 2012. Housing data used in this research is all collected and processed by Chengdu Municipal Urban and Rural Real Estate Administration Bureau. It is basic information of each bargain of commodity housing units from the first quarter in 2005 to the second quarter in 2010. After screening and rejecting of invalid (false) data, the final effective data sample data is 463,518 sets of housing. According to the spatial location, ArcGIS software are input to the 17 sites of first phase of subway Line 1 and 20 sites of the first phase construction of subway Line 2 and all the collected projects of commodity housing to realize specialization, thus to draw distance information from housing to subway stations for further analysis. Fig.1. shows the spatial distribution of housing units and subway stations in Chengdu. Fig. 1. GIS map of Chengdu housing units and subway spatial distribution Variables This paper utilizes hedonic model for regression, the logarithmic form of home price is the dependent variable. This model mainly hopes to observe the different premium brought by different distance from subway site which is one of residen-

5 tial location characteristics. So, in order to get more accurate results, physical characteristic, other location characteristics and other characteristics are controlled. These variables and definitions are listed in Table 1. As to the variable of distance to subway site, different distance dummy variables are separately adopted in different regressions to show the degree of subway effect in different regions and logarithmic values of distance to subway station are used to show the whole developing trend. The physical characteristics of housing consider the area of the housing itself and its floor as well as the whole area and the total floor of the project. As these characteristics have nonlinear effects on housing effects, quadratic term is added to control the nonlinearity, thus to control the housing characteristics from dual aspects of each housing and the whole project. As to location characteristics, block dummy variable is used. As Chengdu is a typical single center city, 26 blocks are classified according to city loop roads and directions. At the same time, location characteristics inside blocks are assumed to have identical quality so as to control location differences among different blocks. Housing sales characteristics and yearly dummy variable are also controlled. Sales characteristics mainly include whether they are existing unites or sales duration, thus controlling effects on home prices from different situations. Yearly dummies mainly control the increase of home prices in different years and effects from the external impacts on home prices in different years (eg. Global financial crisis in 2008). Table 1. Definitions of Variables Category Variable & Definition Dependent variable Home price of unit area (Unit: U.S dollar/m 2 ) Physical Characteristics Location Characteristics Other Characteristics a) Room area of housing unit b) The approved sales area of projects c) Located floor of housing unit d) Total floors of housing unit a) Located block. According to loop and position to classify 26 blocks. (dummy variable) b) Scope. Dummy variables of different spatial scopes around subway stations, such as 0-0.5km, 0.5-1.0km, 1.0-1.5km, 1.5-2.0km, 2.0-2.5km and others. (dummy variable) c) accessibility to subway station--distance between housing unit and subway station a) Sales duration time. The interval between project opening sale time and dealing time b) Whether the unit is completed house or forward house. (dummy variable) c) dealing year of this housing unit. (dummy variable)

6 Empirical Results Analysis of Influenced Scope Five dummy variables with five different distances range are respectively added to the subway influence scope, thus confirming the probable influence scope on surrounding home prices of subway (See Table 2). The regression results show that pricing level of commodity housing distant within 0.5km from subway station is 5.2% higher than those outside 0.5km, those within 1km and 1.5km are 2.2% higher than those outside, those outside 2km are 3% lower than those outside, however, the differences are very little for those inside 2.5km. From results in Table 2, we can see that distance dummies coefficients have an overall declining trend with the widening of radius, and a positive negative turning point within 1.5km-2.0km. The above shows that the spatial effect of Chengdu subway decreases with the increase of distance to subway. The general influence scope is within 2km and it is not obvious or just disappears when exceeding 2km. Table 2. Identify influenced scope of subway stations Dependent variable: ln(home price) Distance to subway station within 0.5km 0.052 *** within 1.0km 0.022 *** (1) (2) (3) (4) (5) within 1.5km 0.022 *** within 2.0km -0.030 *** within 2.5km 0.001 Constant 8.448 *** 8.453 *** 8.446 *** 8.492 *** 8.467 *** Observations 463,518 463,518 463,518 463,518 463,518 R-squared 0.486 0.485 0.485 0.486 0.484 t-value in parentheses; Physical, location and other characteristics of housing unit are all controlled but not reported here; *** p<0.01. As commodity housing may be influenced by other factors, data within 2km are chosen for further study in the following in order to avoid other disturbances while in the process of scrutinizing influence scope (See Table 3). With the same control of other factors influence, equation (1) is a comparison between the commodity housing with the distance of 0-0.5km, 0.5-1km and 1.0-1.5km and the projects within 1.5km-2km; while equation (2) is a logarithm value introduced to subway station distance. And the change gradient of home prices are analyzed directly. The regression results of the above two equations both show that there will be a larger housing value within a closer distance to subway station, of which com-

7 modity housing value is more obvious for those within 0-0.5km. And the commodity housing increment degree is far more than the ordinary housing within 0.5km-2km. Table 3. Scrutinizing influenced scope of subway stations Dependent variable: ln(home price) (1) (2) Housing unit within 0 to 0.5km from the nearest subway station (dummy) Housing unit within 0.5 to 1.0km from the nearest subway station (dummy) Housing unit within 1.0 to 1.5km from the nearest subway station (dummy) 0.140 *** 0.074 *** 0.075 *** Distance to the nearest subway station -0.054 *** Constant 8.377 *** 8.458 *** Observations (within 2km) 223,982 223,982 R-squared 0.532 0.524 Physical, location and other characteristics of housing unit are all controlled but not reported here; *** p<0.01. Estimation of Commodity Housing and Land Premium Chengdu subway Line 1 and 2 obviously bring the premium of the surrounding into commodity housing units nearby stations. The increment of housing within 1.5km of the radius and within 1.5-2km is between 7% and 14%. To estimate conservatively, the increment of the dealt newly built housing is 1.71 billion U.S dollars. If calculated as commodity housing, the increment amount is 5.69 billion. The specific calculation method is as follow: (a)calculate the average price of the newly built commodity housing within 1.5-2km from subway station as the reference value; (b)separately calculate the price appreciation in each spatial scope(0-0.5km, 0.5-1.0km, 1.0-1.5km); (c)with the appreciation in each scope multiplies the total area of selling units, the overall premium in each scope can be known; (d)add these premium of the three scopes, and then the total premium amount of the dealt newly built housing units can be calculated; (e)transform housing premium into residential land premium per unit area, and estimate the total area of residential land in influenced scope; (f)then we can estimate overall premium of residential lands in each scope and whole influenced scope(0-1.5km). The calculation procedures and specific results are shown in Table 4.

8 Table 4. Estimation of commodity housing and land premium Scope (km) Premium percentage (%) Home price on baseline (U.S dollar/m 2 ) Total amount of selling area of housing units (million m 2 ) Housing premium in this scope (billion U.S dollar) 0-0.5 14.0 1141.30 5.53 0.88 0.5-1.0 7.4 1141.30 5.83 0.49 1.0-1.5 7.5 1141.30 3.92 0.34 Total: 1.71 Table 4. (continue) Scope (km) Projection area of housing sample on residential land (thousand m 2 ) Residential land premium per unit area (U.S dollar/m 2 ) Total area of residential land in the influenced scope (million m 2 ) Residential land premium in this scope (billion U.S dollar) 0-0.5 27.99 3157.94 77.6 2.45 0.5-1.0 37.23 1322.86 132.3 1.75 1.0-1.5 25.54 1314.73 113.5 1.49 Total: 5.69 The percentage of residential land in total land area is 30%, the percentage of housing projecting on land in total residential land is 10%. Influence on Home prices from Time Effect Subway s effects on home prices need to take time effect into consideration, which means changes of economy around stations at the key time. The so-called "milestone" refers to a demarcation points like subway s planning and design, working starting, finishing and operation, etc. In view of the operation status of subway Line 1 (till Oct. 2010) and the having been built for 2 years of subway Line 2, beginning date of construction time is quite proper to be the key time point in the analysis of time effect of subway. We classify time period into two parts (before and after work operation), representing the price gradient after work operation and differentials before work operation with cross variables and showing price before working operation change of surrounding land with distance variable coefficient. We find that the ascendant of surrounding home prices along the two subways after work operation is obviously larger than those before operation, which states that subway has an obvious time effect on surrounding home prices. After the work operation of subway, the final location of the station is confirmed, consequently, the future incomes brought by the station is much more riskless. Therefore, more investments will be attracted in the housing market surrounding subway station. As a result, a greater appreciation will be brought. At the same time, more enterprises and businesses will be attracted to have a layout in advance around subway station, thus improving the facilities around subway and bringing

9 grater increment. Therefore, with the improvement of subway information and the strengthening of agglomeration effect, the total premium will realize gradually. Conclusion and Suggestions Governments should fully predict land premium surrounding subway and reasonably adjust time sequence of land grant. Rail transit is a public investment project. Land premium is an important embodiment of its externality and a social profit created by public investment. In theory, there should be some mechanism for value capture, which will also contribute to the recovery of the large amount of capital invested by subway construction. Different systems have different forms as to how to specifically implement the value capture mechanism, and the key is to connect the huge cost of subway construction with the profits brought by premium of the surrounding land. Therefore, before the land premium effect have a panorama (eg. When the subway construction hasn t begun yet), the governments should fully predict the possibility of premium when they grant surrounding lands. Premium should be taken into consideration when assessing base price. Great efforts should be made while granting lands to realize value capture, in order to make up for the subway construction cost. Besides, city governments can also keep the most valuable part of land around the subway. Granting time should be reasonably grasped. This part should be granted after it has a more sufficient increment, so as to better appreciate the profits caused by land appreciation. Combined Development Model should be adopted to relieve government expenditure pressure. Faced with huge subway investment, the government tends to be eager to make up for the construction cost by income of granting land, which may result in a relatively low granting price. Thus the deserved increment profit will lose. According to the experience of Hong Kong Mass Transit Railway, enterprises, with the ability of subway construction and real estate development in the mean time, are found to repair subway and develop the surrounding areas. The whole or parts of costs of subway construction are undertaken by enterprises, and parts of the surrounding areas are granted in a low price (even for free). Development enterprises get profit from the development of the surrounding real estate, which is enough to make up for the cost of subway construction, thus realizing a win-win situation. In the long run, a restitution mechanism for increment profit should be built. If premium effect cannot be fully taken into consideration in the process of subway construction and operation, land premium will always be obtained by individual owner. But as the undertaker for subway construction investment and operation cost, government or subway construction enterprises have less appreciation of these profits. In order to obtain the more matching profits for governments, some countries and districts build a restitution mechanism for increment profit.

10 California State Assembly in 1983 considered that the real estate around subway would get profits from subways. So they authorized the then Rapid Transit Bureau of Southern California to mark Special Advantage Review Tax Area around subway. Special property taxes are collected within these areas and used for repaying the published bonds due to subway construction, which lasts more than 20 years. In our country, similar methods can also be considered to be a system arrangement for premium recycling when system and social conditions are mature. Then each party can have a better coordination and matching between the investment of subway construction and the obtaining profits from subway premium. Acknowledgments This research is supported in part by the National Natural Science Foundation of China (Grant No. 70973065), Major Project of National Social Science Foundation of China (Grant No. 09&ZD042) and Tsinghua University Initiative Scientific Research Program. References [1]Amstrong, R. (1994). Impacts of commuter rail service as reflected in single-family residential property values. Transportation Research Record, 1466: 88-98. [2]Cervero, R., & Duncan, M. (2002). Transit s value-added: Effects of light and commuter rail services on commercial land values. Transportation Research Record, 1805: 8-15. [3]Gatzlaff, D. H., & Smith, M. T. (1993). The impact of the Miami metro rail on the value of residences near station locations. Land Economics, 69(1): 54-66. [4]Gu, Y. Z., & Zheng, S. Q. (2010). The impacts of rail transit on property values and land development intensity: The case of No.13 Line in Beijing. Acta Geographica Sinica, 65(2): 213-223. [5]He, F., & Wang, X. L. (2004). The impacts of the rail transit on property values. China Real Estate, (9): 13-15. [6]Henneberry, J. (1998). Transport investment and house prices. Journal of Property Valuation and Investment, 16(2): 144-158. [7]Knaap, G. J., Ding, C., & Hopkins, L. D. (2001). Do plans matter? The effects of light rail plans on land values in station areas. Journal of Planning Education and Research, 21(1): 32-39. [8]Liang, Q. H., Kong, L. Y., & Deng, W. B. (2007). Impact of URT on real estate value: The case of Beijing Metro Line 13. China Civil Engineering Journal, 40(4): 98-103. [9]Pan, H. X., & Zhong, B. H. (2008). The impact of metro on property price: A case study of Shanghai. Journal of Urban Planning, (2): 62-69. [10]Richard, V. (1993). Changing capitalization of CBD-oriented transportation systems: Evidence from Philadelphia, 1970-1988. Journal of Urban Economics, 33(3): 361-376. [11]Zheng, J. F., & Liu, H. Y. (2005). The impact of URRT on house prices in Shenzhen. Journal of the China railway Society, 27(5): 11-18.