1# RECORD f90cf5207d8f2a48c20e426cb00e7a3d92fecdbe2# TITLE: The Capitalization Effect of Natural Amenities on Housing Price in Urban China: New Evidence From Changsha3# YEAR: 2022.0 VENUE: Frontiers in Environmental Science DOI: 10.3389/fenvs.2022.8338314# AUTHORS: Tuo-Fu Huang; Qing-Yun He; Xiao Ouyang5# PAGES: 16 CANDIDATE TABLE PAGES (1-based): [8, 9]6# ABSTRACT: This study takes Changsha, an emerging mega-city in China, as a sample to analyze the influence of natural amenity (NA) on housing prices during different periods. We divided the NA into four categories according to their characteristics: Green Space (GS), Wetland Park (WP), XiangRiver Scenic Belt (XR), and green ratio within a community. The housing transaction data is used to calculate the accessibility of a settlement to nearest NA via the distance. The capitalization effect of NA is investigated through a geographic weighted regression (GWR), combined with the hedonic theory. Results show that ①the impact and scope of NA on housing prices gradually expanded during years. ②As the key nature landscape of Changsha, the impact coefficient of XR has risen from 2.34% in 2012 to 5.32% in 2020. WP has a remarkable increase for affecting housing prices, rising from 1.24 to 3.65%. Effect of GS keeps in line during years, at 2.34–3.72%. And green ratio’s impact increase from 3.61 to 4.55%. ③The higher capitalization effect of NA undergoes a spatial expansion from the central area to urban fringe over time, which also implies the capitalization of NA is becoming more pronounced. ④Residential communities impacted by NA have a propensity to separate over time. This study provides a different perspective in defining the importance of natural amenities in urban habitat of modern China.789===== PAGE 1 (score 13, numeric-paren 0, table-words 0, var-words 12) =====1011The Capitalization Effect of Natural12Amenities on Housing Price in Urban13China: New Evidence From Changsha14Huang Tuofu 1, He Qingyun 1* and Ouyang Xiao 2*151College of Geographic Science, Hunan Normal University, Changsha, China, 2Hunan Institute of Economic Geography, Hunan16University of Finance and Economics, Changsha, China17This study takes Changsha, an emerging mega-city in China, as a sample to analyze the18influence of natural amenity (NA) on housing prices during different periods. We divided the19NA into four categories according to their characteristics: Green Space (GS), Wetland Park20(WP), XiangRiver Scenic Belt (XR), and green ratio within a community. The housing21transaction data is used to calculate the accessibility of a settlement to nearest NA via the22distance. The capitalization effect of NA is investigated through a geographic weighted23regression (GWR), combined with the hedonic theory. Results show that ①the impact and24scope of NA on housing prices gradually expanded during years. ②As the key nature25landscape of Changsha, the impact coefficient of XR has risen from 2.34% in 2012 to265.32% in 2020. WP has a remarkable increase for affecting housing prices, rising from 1.2427to 3.65%. Effect of GS keeps in line during years, at 2.34–3.72%. And green ratio’s impact28increase from 3.61 to 4.55%. ③The higher capitalization effect of NA undergoes a spatial29expansion from the central area to urban fringe over time, which also implies the30capitalization31of32NA33is34becoming more pronounced.35④Residential36communities37impacted by NA have a propensity to separate over time. This study provides a38different perspective in defining the importance of natural amenities in urban habitat of39modern China.40Keywords: natural amenities, hedonic theory, GWR, spatial expansion, Changsha city41INTRODUCTION42China has experienced a world-renowned urbanization process, with the rate growing from 16.22%43in 1978 to 63.89% in 2020, and the Chinese urban landscape has changed tremendously (Yibo et al.,442020; Alix-Garcia et al., 2014). Rapid urbanization and the associated increase in vehicle traffic have45many negative implications, including increased air and water pollution, noise and congestion.46Physiological ailments such as chronic bronchitis and pneumonia, as well as mental problems as47depression and burnout syndrome, have emerged as a significant threat to public health. However,48urban settings also have features that promote health and may assist in attenuating some of the49negative health implications of urban life. Natural amenities (NA) is increasingly being viewed as an50essential settings for health promotion to solve city residential problems (De Vries et al., 2003;51Anderson and West, 2006; Rajapaksa et al., 2017). The past decades, in particular, have seen an52incline in interest for ecological, nature-based health promotion initiatives in both studies and53practices. The new strategy of urban planning is to create, conserve, maintain, and develop NA across54urban areas (Gómez-Baggethun and Barton, 2013; Tuofu et al., 2021).55Edited by:56Bao-Jie He,57Chongqing University, China58Reviewed by:59Junsong Wang,60South China University of Technology,61China62Jun Yang,63Northeastern University, China64*Correspondence:65He Qingyun66heqingyun1848@gmail.com67Ouyang Xiao68xiao.ouyang@foxmail.com69Specialty section:70This article was submitted to71Land Use Dynamics,72a section of the journal73Frontiers in Environmental Science74Received: 12 December 202175Accepted: 03 February 202276Published: 28 February 202277Citation:78Tuofu H, Qingyun H and Xiao O (2022)79The Capitalization Effect of Natural80Amenities on Housing Price in Urban81China: New Evidence From Changsha.82Front. Environ. Sci. 10:833831.83doi: 10.3389/fenvs.2022.83383184Frontiers in Environmental Science | www.frontiersin.org85February 2022 | Volume 10 | Article 83383186187ORIGINAL RESEARCH88published: 28 February 202289doi: 10.3389/fenvs.2022.833831909192===== PAGE 8 (score 18, numeric-paren 0, table-words 3, var-words 10) =====9394studies. We are primarily concerned with the estimated values95and trends of NA variables.96Among the many structural factors influencing a property’s97pricing, the green ratio has a long-term and growing impact. The98house price can rise by 3.61% in 2012 and 4.38% in 2020 for each99level increase in the green ratio. Moreover, the GWR’s estimate100rate is increased even faster, from 3.04% in 2012 to 4.55% in 2020.101Evidence shows that the requirements for nature amenities within102a community of home buyers in Changsha have increased rapidly103in the last decade.104The influence of proximity to WP, XR and GS on housing105price have progressively risen, with varying rates from OLS. XR106had the most significant effect on housing prices in 2012 of all107NA. Each additional kilometer of distance resulted in a 3.45%108decline in housing values. At the time, wetlands had a coefficient109of 1.17%, while green spaces had a coefficient of 1.98%. In 2016,110XR effect coefficient climbed slightly to 3.65%, but wetlands and111green spaces influence on values increased more rapidly, by 1.76%112t and 2.67%, respectively. WP have the most excellent effect on113values in 2020, with a coefficient of 3.34%, followed by XR and114GS, with coefficients of 4.08 and 3.28%, respectively. GWR mean115results of wetland, XR and green spaces have the same trend, from1162012 at 1.24, 2.34, and 2.75–3.65, 5.32, and 3.72% in 2020,117respectively. The increased capitalization of the three forms of118NA accessibility also shows the growing awareness of residents119toward120the121natural122environment123and124green125recreational126opportunities in the immediate vicinity. Simultaneously, the127mean GWR values for all three accessibility factors are higher128than OLS coefficients, indicating that the sample variables are129more susceptible to the effect of housing prices.130With respect to other key variables, the capitalization effects131follow a different trend. CBD still determines housing prices132significantly, although this influence is gradually diminishing.133Each kilometer away from the CBD Wuyi square influenced1346.91% of samples’ housing prices in 2012, compared with 5.82135and 4.61% in 2016 and 2020. In GWR mean view, the results are136at 8.41, 8.15, and 6.26% during the same time. The structure of137Changsha city is rapidly changing, as evidenced by the fact that138housing prices are decreasingly dependent on the accessibility of139the city’s core. A dispersed polycentric pattern has emerged as140significant NA around housing and residency are accumulating in141other142locations143outside144the145core146area.147Neighborhood148characteristics such as metro and educational institutions also149have an effect on housing prices. Due to the shortage of150educational151facilities152in153China’s154main155cities,156property157ownership has become a popular market-based method of158acquiring high-quality schools. Accordingly, the impact factor159for quality middle schools in Changsha with GWR has increased160from 3.24 to 5.72%. The metro is a major transportation route in161modern China, often increasing the value of surrounding real162estate. In Changsha, this impact on property prices is from 3.78%163in 2012 to 4.24% in 2020. The rest variable estimates are shown164with Table 4.165It was easily found that statistic test values of GWR were166higher for all 3 years after accounting for the spatial heterogeneity167variables, indicating the advantage of GWR more suitable to deal168with spatial heterogeneity. As shown in Table 5, while the R2169values of all models have fallen over time, indicating that the170variables171impacting172housing173prices174are175becoming176more177complicated. However, comparing the OLS model with the178GWR model, we can see that the fitted value of GWR is much179better than that of OLS, and the residual squares and standard180deviations of GWR are significantly lower than those of OLS.181Therefore, the GWR model fitting effect is superior to the general182OLS regression effect, making it a suitable approach for analyzing183the evolution of the spatial structure of residential pricing in184Changsha and the evolution of heterogeneity of crucial factors.185The Spatial Change of Estimates186The spatial distribution of the estimated coefficients for the years1872012, 2016 and 2020 are compared to show the trend of variables.188Since we focus on NA, the variables XR, WP, GS and green ratio189are the main objects to study. Simultaneously, due to the190accessibility to the CBD being the primary factor impacting191the urban structure, we also analyze this variable. To facilitate192comparison, we group the coefficient values into five average193quantiles, each one covering a 20% interval in order, from the194lowest to the highest.195The distribution of GWR estimate value of CBD variable is196shown in Figure 5. The distance between the house sample and197Changsha’s CBD Wuyi Square has a negative effect on the198housing prices. In terms of the spatial and temporal evolution199of the coefficient values, their spatial distribution changes more200noticeably and exhibits a trend throughout the three cross-201sectional years. In 2012, the CBD variable had a high value in202the north of Changsha’s urban core area, notably along the203XiangRiver’s north coast and in the central section of Kaifu204District, with a coefficient value of lower than −29% per km.205The area of a strong influence of CBD variable in 2012 is relatively206small, while the impact value is large. In 2016, the high-value area207of CBD expanded into the center of the city, including a part of208Yuhua and Tianxin districts, the northern part of Yuelu district.209Although the area has expanded, the impact values have210decreased, with the value is greater than −20% per km. In2112020, the CBD high-value area was mainly around Wuyi CBD212and Changsha south region, while the northern part of the city,213west of XR, is less severely affected. In general, the high-value214region of the CBD variable expands and then contracts,215progressively shifting the impact area from the north-central216to the southern section of Changsha city. This change indicates217that as the overall size of the city grows and polycentric structure218emerges, the correlation between CBD accessibility, as an219indicator,220and221housing222prices223tends224to225decline.226The227diminishing228influence229of230the231CBD232on233price234while235the236increasing impact of NA accessibility on price is a sample of237the evolution toward polycentric city structure. It could be238implied that the development of NA is one of the driving239forces in the evolution of the urban spatial structure of Changsha.240The distribution of GWR estimate value of WP is shown in241Figure 6. With the rapid development of wetlands amenities in242Changsha, an increasing number of people are choosing to locate243homes close to them, attracting more housing development with244high-density. Therefore, the capitalization effect of NA on245housing price increases as market develops, and this effect246Frontiers in Environmental Science | www.frontiersin.org247February 2022 | Volume 10 | Article 8338312488249Tuofu et al.250Capitalization Effect of Natural Amenities251252253===== PAGE 9 (score 56, numeric-paren 30, table-words 4, var-words 3) =====254255reflects spatial heterogeneity over time. In 2012, the high values256for the variable area of wetlands were mainly in the southern part257of core city, with the high value of the local capitalization258coefficient being between 3 and 7%.In 2016, the range of WP’s259high impacts changed to the east of the city, and the coefficient260topped 12%, higher than in 2012. House prices in Changsha261TABLE 4 | 2012, 2016, 2020 Global regression results and GWR mean results.262Year 2012263Year 2016264Year 2020265Variable266Estimate267t value268Mean269Estimate270t value271Mean272Estimate273t value274Mean275Cons2768.855364277736.10382788.7521632798.839498280830.02952818.8410372828.897656283546.63612848.868051285Plotratio2860.02112*2871.5657062880.0327612890.022562900.1914912910.01054292−0.02328**293−1.70085294−0.01448295Greenratio2960.03613***2974.056152980.0304352990.02976***3004.142,8143010.0482893020.04382***3035.7862373040.045568305P.M3060.01167**3071.8346223080.0194063090.02562***3102.3499363110.0314753120.04145***3132.3238113140.044331315B.Age316−0.01512*317−1.54565318−0.047719319−0.01722**320−1.69759321−0.03906322−0.01691***323−3.53801324−0.02237325Subway3260.06076***3274.508373280.0378993290.05351***3304.2565873310.0438363320.05325***3332.898673340.042437335QP.school3360.02421*3371.350063380.0076323390.0225023400.8361273410.0221313420.0293943430.1157193440.015944345QM.school3460.01405**3471.7666993480.032423490.030255*3501.5198393510.0389273520.05389***3532.3804893540.05728355d.CBD356−0.06916***357−3.71642358-0.08411359−0.05889***360−3.41646361−0.08158362−0.04623***363−3.15669364−0.06261365d.Wetland366−0.01178***367−2.87305368−0.01241369−0.01763***370−2.42941371−0.02419372−0.03341***373−3.11678374−0.03645375d.Xriver376−0.03445***377−2.7655378−0.02342379−0.03656***380−2.4744381−0.02694382−0.04085***383−2.52188384−0.05321385d.Gspace386−0.01986***387−2.22424388−0.02752389−0.02671***390−3.01611391−0.03334392−0.03286***393−2.36141394−0.03727395N3965133976793981,013399TABLE 5 | Comparison test parameters of OLS model and GWR model for 2012, 2016, 2020.400201240120164022020403Test model404OLS405GWR406OLS407GWR408OLS409GWR410Classic AIC411219.39489412148.154552413200.650388414181.790936415233.869488416192.747132417AICc418220.627322419185.185672420201.810693421186.910349422235.733249423201.089886424BIC/MDL425297.742219426556.019445427286.429929428361.450754429305.074561430341.377132431CV4320.0856754330.0935114340.0791554350.07774443613.9021624379.855557438R24390.1753084400.440464410.1923134420.2614994430.2991474440.430731445Adjusted R24460.1500934470.2990764480.1701514490.2046944500.2667714510.355767452Residual Ssquares45346.25898145431.38599145550.28964445645.98183845737.73452945830.649917459FIGURE 5 | Spatial and temporal variation of CBD influence coefficients in 2012, 2016 and 2020.460Frontiers in Environmental Science | www.frontiersin.org461February 2022 | Volume 10 | Article 8338314629463Tuofu et al.464Capitalization Effect of Natural Amenities465466467===== PAGE 10 (score 10, numeric-paren 0, table-words 0, var-words 10) =====468469increased sharply from 2017 to 2020, with the extreme value of470the coefficient exceeding 20% from 2020, much higher than in4712016. During this period, several well-known wetlands in Yuelu472District, such as Meixi Lake and Yanghu Lake, were highly sought473after by homebuyers and became a rallying point for high-priced474houses, so the influence of blue on surrounding house prices475increased significantly. The area of coefficient extremes is476concentrated in Meixi Lake, Yanghu Lake, and the city’s east,477outside the core urban area, which may imply that homebuyers478are increasingly attracted to fringe blue space. Nonetheless, Yuelu479district as a significant urban ecological region in Changsha has480been enhanced, affecting housing prices.481XR is the traditional center of Changsha’s natural landscape,482and its accessibility has a significant impact on housing prices.483The spatial heterogeneity variation of this effect was found via484GWR during times, as shown by Figure 7. In 2012, the main485impact high value area of the variable radiated is the southern486area of Yuelu District outside the western ring road, with the487highest coefficient value over −20%. During this period, Yanghu488New Town has been gradually developed, and its main high-end489properties are located along XR, which is typical of river-view490buildings. At this time, on the east bank of the river, lots of older491houses with poorer scenic, which are not as sensitive to distance492to this river as high-end river view houses. Only few new493buildings high-rise residences along the riverbank that are494more affected by the landscape of XR in this region. In 2016,495the high value area spread further to the east of the river and south496of the city as it expanded southward with the construction of a497large498number499of500high-rise501residential502buildings.503The504concentration of high-system value properties in the southern505region increased significantly, but the impact high-value diminish506to about −15%.507In 2020, the extreme value further decreased to about −10%.508With the rise of a new residential area on the north side of the509west bank of XiangRiver and the development of areas across the510river, the high-value area of the access variables is further511expanded. Overall, the residential neighborhoods whose house512prices are affected by XR accessibility variable are more robust in513the west of the river than in the east, and there is a higher density.514Despite residential samples in the newly developed area being515more affected by XiangRiver, the regression coefficients of the5163 years have decayed, indicating that the real influence of this517accessibility variable has weakened.518The green ratio is an important NA within a community, and519it has a profound impact on the housing price of the community.520This impact also has very obvious temporal and spatial521differentiation as shown in Figure 8. In 2012, the extreme522impact area of greening rate on house price was mainly523distributed in the core area within the east ring road of the524city as well the south city area, with the high value range between5255 and 6%.Most of the development areas of good quality526properties in this period were located along with XR in the527core of the city, which affected the degree of community528greening.529Simultaneously,530there531is532a533tendency534that535the536surrounding NA on the east ring road is lower than in west,537making property purchasers more sensitive to the NA inside538community. 2016, high-value area expanded to the west bank of539XR and the east part of Changsha, while the impact of the south540city was significantly declined. In 2020, the number of properties541covered by the high-value affection in east XR was significantly542reduced. The high-value areas are mainly concentrated in several543large residential groups of Meixi, Yanghu Lake and north region,544with a decreasing distribution pattern of coefficient values from545north to south. The average value of the estimated coefficient for546those 3 years has increased gradually, which is maintained547between 6 and 10%.548As shown in Figure 9, the impact of green space on residential549prices is broadly distributed in urban Changsha, and the high550FIGURE 6 | Spatial and temporal variation of WP influenced coefficients in 2012, 2016 and 2020.551Frontiers in Environmental Science | www.frontiersin.org552February 2022 | Volume 10 | Article 83383155310554Tuofu et al.555Capitalization Effect of Natural Amenities556