1# RECORD bd42cea5779904b39c488f5e59235573696a28272# TITLE: THE IMPACT OF URBAN RENEWAL ON NEIGHBORING HOUSING PRICES: AN APPLICATION OF HIERARCHICAL LINEAR MODELING3# YEAR: 2022.0 VENUE: International Journal of Strategic Property Management DOI: 10.3846/ijspm.2022.159714# AUTHORS: Chun-Chang Lee; Chih-Min Liang; Wen-Chih Yeh; Zheng Yu5# PAGES: 13 CANDIDATE TABLE PAGES (1-based): [9, 10]6# ABSTRACT: This study explored the impacts of urban renewal projects on neighboring housing prices. Hierarchical linear modeling (HLM) was employed to analyze urban renewal projects in Taipei City. The Level 1 independent variables pertained to a house itself (19,157 pieces of data), such as its structure and neighborhood attributes. The Level 2 variable pertained to an urban renewal project (23 cases of urban renewal), and the explanatory variable was the scale of each urban renewal project. The study examined whether differences exist between the impacts of various urban renewal projects on neighboring housing prices, and analyzed the extent to which the differences in neighboring housing prices are caused by the differences between urban renewal projects. The empirical results showed that the mean housing price varies significantly between each urban renewal project. In regard to the variance in the mean house price, 31.46% was caused by the differences between the urban renewal projects. The estimated coefficient of the grand floor area of urban renewal (FLAREA) had a positive value and attained a 1% level of significance. This indicates that the larger the scale of an urban renewal project, the larger its effects on neighboring housing prices. The empirical results of this study could better explain the impacts of the scale of an urban renewal project on the externalities of urban renewal.789===== PAGE 1 (score 14, numeric-paren 1, table-words 1, var-words 11) =====1011Copyright © 2022 The Author(s). Published by Vilnius Gediminas Technical University12This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unre-13stricted use, distribution, and reproduction in any medium, provided the original author and source are credited.14Introduction15According to Article 4 of Taiwan’s Urban Renewal Act, 16the primary objectives of urban renewal are to promote 17comprehensive environmental sustainability, improve the 18quality of living, and transform the appearance and aes-19thetics of a city. The process of urban renewal entails three 20approaches: reconstruction, renovation, and maintenance. 21In reconstruction, specific urban renewal areas are desig-22nated where the existence of worn-out houses could po-23tentially impact their surrounding area, and these build-24ings are demolished on a large scale in order to enhance 25the quality of living and to promote the growth of hous-26ing prices in neighboring areas. In renovation and main-27tenance, houses that have a dilapidated exterior, a visually 28unappealing appearance, or a deteriorated functionality 29are subjected to small-scale renovations and facelifts to 30enhance their functionality and usage as well as to beau-31tify the city they are found in.32With regard to the issue of urban renewal on housing 33prices, many studies have explored the impacts of urban 34renewal laws and provisions and analyzed the imple-35mentation of urban renewal projects. For instance, Jones 36(2008) and Hyra (2008) examined urban renewal and the 37post-war “new urban renewal” in the United States. Both 38authors focused on the racial differences and expounded 39unjust urban renewal projects by attaching importance to 40the process of community participation and the continu-41ous development of housing prices. Some studies and data 42reviewers had analyzed the impacts of an urban renewal 43project on housing prices within the project area after its 44completion, and few had studied the impacts of urban re-45newal and reconstruction on housing prices in neighbor-46ing areas. In a recent study, Lee et al. (2017) investigated 47the impacts of urban renewal delineation time on housing 48prices in neighboring areas by employing the difference-49in-difference estimation approach. The empirical results 50International Journal of Strategic Property Management51ISSN: 1648-715X / eISSN: 1648-9179522022 Volume 26 Issue 1: 11–2353https://doi.org/10.3846/ijspm.2022.1597154*Corresponding author. E-mail: lcc@mail.nptu.edu.tw55THE IMPACT OF URBAN RENEWAL ON NEIGHBORING HOUSING 56PRICES: AN APPLICATION OF HIERARCHICAL LINEAR MODELING57Chun-Chang LEE1,*, Chih-Min LIANG2, Wen-Chih YEH3, Zheng YU4581 Department of Real Estate Management, National Pingtung University, Pingtung, Taiwan592 Department of Public Finance and Tax Administration, National Taipei University of Business, Taipei, Taiwan603 Department of Real Estate Management, HungKuo Delin University of Technology, New Taipei, Taiwan614 Department of Land Economics, National Chengchi University, Taipei, Taiwan62Received 24 May 2021; accepted 10 September 2021; first published online 17 January 202263Abstract. This study explored the impacts of urban renewal projects on neighboring housing prices. Hierarchical linear 64modeling (HLM) was employed to analyze urban renewal projects in Taipei City. The Level 1 independent variables per-65tained to a house itself (19,157 pieces of data), such as its structure and neighborhood attributes. The Level 2 variable 66pertained to an urban renewal project (23 cases of urban renewal), and the explanatory variable was the scale of each 67urban renewal project. The study examined whether differences exist between the impacts of various urban renewal pro-68jects on neighboring housing prices, and analyzed the extent to which the differences in neighboring housing prices are 69caused by the differences between urban renewal projects. The empirical results showed that the mean housing price varies 70significantly between each urban renewal project. In regard to the variance in the mean house price, 31.46% was caused 71by the differences between the urban renewal projects. The estimated coefficient of the grand floor area of urban renewal 72(FLAREA) had a positive value and attained a 1% level of significance. This indicates that the larger the scale of an urban 73renewal project, the larger its effects on neighboring housing prices. The empirical results of this study could better explain 74the impacts of the scale of an urban renewal project on the externalities of urban renewal.75Keywords: urban renewal, housing prices, grand floor area, scale of an urban renewal project, hierarchical linear modelling.767778===== PAGE 9 (score 33, numeric-paren 2, table-words 5, var-words 14) =====7980International Journal of Strategic Property Management, 2022, 26(1): 11–23811982on urban renewal, in order to prevent overextension of the 83distance to affect the estimation. Therefore, we narrowed 84the range of effects to within 800 meters. On this basis, 85800 meters served as the comparison benchmark of the 86range of influence in this study, and data of houses located 87outside the range of influence was removed.88To measure the actual distance from a house to pub-89lic infrastructure, this study utilized a geographic infor-90mation system software – Arc GIS version 10.1 to link 91the data of a house’s internal and external characteristics, 92while integrating the spatial location data of the house and 93the sample image layer data to construct a database on a 94house’s spatial attributes. This study obtained the address-95es of transacted houses from the Taiwan Real Estate Trans-96action Bulletin. Afterwards, Google Maps and the Taiwan 97Geospatial One-Stop Portal (TGOS) were employed to 98convert all sets of house transaction data into spatial data 99in accordance with the addresses of transacted houses ob-100tained from the Taiwan Real Estate Transaction Bulletin. 101Afterwards, the location of the sample points of the house 102transaction spatial data was established in the image layer 103of Taipei City’s administrative districts. After constructing 104each set of spatial data on housing transactions within the 105scope of research, the geographic information system soft-106ware Arc GIS version 10.1 developed by Environmental 107Systems Research Institute, Inc. (ESRI) was used to overlay 108the layer of the location of the house transaction spatial 109data on other layers of relevant public infrastructure, such 110as MRT stations, railway stations, and schools, with the 111purpose of calculating the nearest distance from each spa-112tialized housing data to public infrastructure.1134. Descriptive statistics of samples114The descriptive statistics of the sample is presented in Ta-115ble 2. The mean housing price was NT$26.6933 million. 116(1 USD = 27.9 NTD, January 6, 2021) The mean housing 117area was 43.28 ping. The mean house age was 21.90 years. 118The mean number of rooms was 2.92. The mean number 119of living rooms was 1.85. The mean number of bathrooms 120was 1.74. 7% of houses were located on the first floor. In 121terms of the type of housing, 44%, 30%, and 26% of houses 122were high residential buildings, low residential buildings, 123and apartment buildings, respectively. 37% of houses had 124a parking lot(s). The mean distance to the city center was 125598.68 m. The mean distance to the nearest elementary 126school was 365.74 m. The mean distance to the nearest 127junior high school was 503.63 m. The mean distance to 128the nearest MRT station was 632.41 m. The mean distance 129to the nearest railway station was 3276.43 m. The mean 130grand floor area of urban renewal was 5404.24 m2.1315. Empirical results132First, we used the null model to test the variation in the 133mean house price between various urban renewal projects. 134As shown in Table 3, with regard to fixed effects, the es-135timated value of the mean house price g00 is 12.095; with 136regard to random effects, the estimate value of t00 is 0.028, 137and the statistics of c2 is 8861.028, the degree of freedom 138is 22 (J = 23 projects), attaining a 1% level of significance. 139These results show that the mean housing price varies sig-140nificantly between each urban renewal project.141The variance t00 of the mean urban renewal project 142price b0j and the individual level intra-group variance s2 143can be used to calculate the intra-group correlation coef-144ficient (ICC), which is 0.3146 (0.028 / (0.028 + 0.061) = 1450.3146). According to Cohen’s (1998) recommendations, 146the correlation is strong. This indicates that 31.46% of the 147variance in the mean house price is caused by the differ-148ences between the urban renewal projects. Thus, it is im-149portant to take into account the differences between urban 150renewal projects.151Table 2. Descriptive statistics of the sample152Variable153Mean154S.D.155Minimum Maximum156Level 1157PRICE (in 158NT$10,000 units)1592669.331601676.6416133.5016224500.00163AREA (ping)16443.2816520.0716620.00167149.83168AGE (years)16921.9017013.471711.0017249.98173ROOM1742.921750.8317611776178LIVROOM1791.851800.4018111823183BATH1841.741850.5918611875188FLOOR11890.07190–19101921193BUILDING1940.44195–19601971198HUAXIA1990.30200–20102021203PARKING2040.3720502061207DISTURBA208598.68209261.402100211999.99212PRIM213365.74214173.1621502161114.77217JUNI218503.63219255.852203.362211436.33222MRT223632.41224482.352253.702262388.18227TRAIN2283276.432292474.5823068.962318842.66232Level 2233FLAREA2345404.242354274.29236906.6223717669.32238Table 3. Null model analysis results239Fixed effects240Coefficients241Standardi242zed errors243t-ratio244p-value245g0024612.095***2470.035248344.3862490.001250Random 251effects252Variance 253components254df255Chi-256square257p-value258t002590.028***260222618861.0282620.001263Level-1 s22640.061265Deviance 266(-2LL)267933.978268Number of 269estimated 270parameters2712272Note: * indicates p < 0.1, ** indicates p < 0.05, *** indicates p < 0.01.273274275===== PAGE 10 (score 75, numeric-paren 17, table-words 14, var-words 24) =====27627720278C.-C. Lee et al. The impact of urban renewal on neighboring housing prices: an application...279The empirical estimation results are presented in Ta-280ble 4. The estimated coefficient of the grand floor area of 281urban renewal (FLAREA) was 0.001 and attained a 1% 282level of significance. This indicates that the large the scale 283of an urban renewal project, the larger its effects on neigh-284boring housing prices. Guidry et al. (1999) explored the 285effects of urban residential land size on residential land 286prices. The results demonstrated that lot size had a posi-287tive effect on land price, albeit not statistically significant. 288With regard to the housing structure attributes, the sym-289bols of most of the variables were as expected. The es-290timated coefficient of house area was 0.001 and attained 291a 1% level of significance. This indicates that the larger 292the house area, the higher the price. Freybote and Fruits 293(2015) demonstrated that house area has a significant and 294positive effect on housing prices. The estimated coefficient 295of building age and the square of building age was –0.012 296and 0.001, respectively, and both coefficients attained a 2971% level of significance. This indicates that the higher 298the building age, the lower the price, and the smaller the 299price drop. Wilhelmsson (2008) also employed house age 300and the square of house age as explanatory variables, and 301showed the negative impacts of house age, in which the 302impacts decrease with increasing house age. The estimated 303coefficient of the number of living rooms was 0.024, and 304attained a 1% level of significance. This indicates that the 305higher the number of living rooms, the higher the hous-306ing price. Huang et al. (2010) stated that a living room is 307an internal layout of a house and the more living rooms 308a house has, the higher its floor space as well as its price. 309They expected that the number of living rooms have a 310positive impact on housing price as well. The estimated 311coefficient of the number of bathrooms was 0.049, and 312attained a 1% level of significance. This indicates that the 313higher the number of bathrooms, the higher the hous-314ing price. Baltagi and Bresson (2011) demonstrated the 315positive influence of the number of bathrooms on housing 316price. The estimated coefficient of floor number was 0.167, 317and attained a 1% level of significance. This indicates that 318the mean price of houses on the first floor was 16.7% high-319er than those that are not on the first floor. Houses on 320the first floor can also be used for commercial purposes, 321which explains why they often have higher prices. The es-322timated coefficients of high residential buildings and low 323residential buildings was 0.197 and 0.174, respectively, 324and coefficients attained a 1% level of significance. This 325indicates that the mean price of high residential buildings 326and low residential buildings was 19.7% and 17.4% high-327er than apartment buildings, respectively. Since high and 328low residential buildings have many floors, due to safety 329concerns, housing prices would increase in response to 330higher construction costs. The estimated coefficient of the 331number of rooms was –0.016, and attained a 1% level of 332significance. The estimated coefficient of the availability of 333parking lots was –0.065 and attained a 1% level of signifi-334cance. Christiansen et al. (2017) highlighted that houses 335with parking lots would pose a higher challenge to afford-336able housing. This also shows that for house buyers who 337purchased low- to mid-priced houses, buying a house with 338a parking lot increases affordability, and they would rather 339buy houses located near a MRT station or those with a 340larger floor space.341With regard to the neighborhood environment attrib-342utes, the estimated coefficient of the distance to the city 343center –0.001 and attained a 1% level of significance. This 344indicates that the closer a house is to the city center, the 345higher the price. Sunak and Madlener (2016) pointed out 346that the city center offers more infrastructure, is more 347developed, and has more commercial and business ben-348efits. Therefore, the closer a house is to the city center, 349the higher its price. The coefficient of the distance to the 350nearest elementary school (PRIM) was 0.001 and attained 351a 1% level of significance. The coefficient of the distance 352to the nearest junior high school (JUNI) was –0.001 and 353attained a 10% level of significance. This indicates that a 354junior high school offers school district-related benefits to 355houses nearby. Proximity to a school not only ensures the 356accessibility and safety of school children, but also increas-357es recreational space for residents. Therefore, house buyers 358often consider whether a house is located near a school 359when they make their purchases. The estimated coefficient 360of the distance to the nearest MRT station was –0.001 and 361attained a 1% level of significance. This indicates that the 362closer a house is to a MRT station, the higher the price. 363Many middle-to-low income groups rely on public trans-364portation to commute or travel, and hence a house that is 365located nearer to a MRT station is often more expensive 366(Ahlfeldt & Maennig, 2010). The estimated coefficient of 367the distance to the nearest railway station (TRAIN) was 3680.001, and attained the 1% level of significance. Nelson 369Table 4. Empirical estimation results370Fixed effects371Coefficients372Standardized 373errors374t-ratio375g0037611.960***3770.0053782203.893379FLAREA380g013810.001***3820.0013838.803384AREA385g103860.001***3870.0013886.831389AGE390g20391–0.012***3920.001393–18.253394AGES395g303960.0002***3970.00139818.743399ROOM400g40401–0.016***4020.003403–5.233404LIVROOM405g504060.024***4070.0054084.312409BATH410g604110.049***4120.00541310.971414FLOOR1415g704160.167***4170.00841820.975419BUILDING420g804210.197***4220.00742329.409424HUAXIA425g904260.174***4270.00642828.168429PARKING430g100431–0.065***4320.006433–10.403434DISURBA435g110436–0.001***4370.001438–6.099439PRIM440g1204410.001***4420.00144310.814444JUNI445g130446–0.001*4470.001448–1.775449MRT450g140451–0.001***4520.001453–39.312454TRAIN455g1504560.001***4570.0014585.638459Note: * indicates p < 0.1, ** indicates p < 0.05, *** indicates p < 0.01.460461462===== PAGE 11 (score 11, numeric-paren 0, table-words 1, var-words 9) =====463464International Journal of Strategic Property Management, 2022, 26(1): 11–2346521466(2004) and Simons and Jaouhari (2004) concurred that 467while proximity to a railway station offers ease of transpor-468tation, a house that is too close to a railway station would 469be exposed to high noise levels and a high frequency of 470accidents, which has negative impacts on its price. Comber 471and Arribas-Bel (2017) pointed out that crime rates could 472be higher in places near railway stations.473Conclusions and recommendations474Conventional hedonic pricing models generally use the 475ordinary least squares (OLS) estimation approach, spatial 476econometrics or other methods, in which data with dif-477ferent hierarchical characteristics are confined to a single 478level for data processing, and the estimation of hierarchi-479cal problems are not taken into account. Since housing 480units are nested within various urban renewal projects, 481the impacts of an urban renewal project on neighboring 482housing prices should take into account the intrinsic dif-483ferences of the project on neighboring housing prices. Hi-484erarchical linear modeling (HLM) was employed for data 485analysis. Compared to the conventional ordinary least 486squares (OLS) models, HLM can further provide deeper 487insights into the relationships between housing prices and 488the characteristics of urban renewal projects.489Arribas et al. (2016) think that OLS couldn’t solve the 490within-group correlation of datasets on housing prices or 491the correlation between housing prices and neighborhood 492environment. Housing information belongs to the nested 493structure. If we use single-level OLS, it will lead to infer-494ence errors. In past research, the paper by Lee et al. (2017) 495and Liang et al. (2020) adopted HLM model to study the 496effects of urban renewal on housing prices, and Liang et al. 497(2020) adopted spatial econometrics for their research. 498They both neglected the hierarchical structure of hous-499ing information. This will generate bias in the modeling 500estimation. In addition, it will not be possible to observe 501the cross-level influence of the effects of urban renewal. In 502this study, the HLM we adopted considered hierarchical 503structures, which can provide more information for the 504estimation and the result will be more accurate (Arribas 505et al., 2016). By using HLM analysis, we can understand 506how variations in individual houses affect Level 1 and how 507(the explained variation) urban renewal affects Level 2.508The empirical results showed that the mean housing 509price varies significantly between each urban renewal 510project. 31.46% of the variance in the mean house price is 511caused by the differences between the urban renewal proj-512ects. In this regard, it is important to take into account the 513differences between urban renewal projects when examin-514ing the impacts of urban renewal projects on neighboring 515housing prices. The estimated coefficient of the grand floor 516area of urban renewal (FLAREA) had a positive value and 517attained a 1% level of significance. This indicates that the 518larger the scale of an urban renewal project, the larger its 519effects on neighboring housing prices. Our findings show 520that urban renewal projects of different scales have differ-521ent impacts on neighboring housing prices. Previously, we 522also used 500 meters as the range of effects for estimation, 523but there was no significant difference in the results.524Subsequent research could include other features of 525urban renewal for analysis to allow for a more thorough 526analysis on the impacts of urban renewal on neighbor-527ing housing prices. In addition, we suggest examining 528whether the features of urban renewal projects moderate 529the impacts of housing features on housing prices, that is, 530whether moderating effects exist between the aggregate-531level variables and the individual-level variables.532Level 1 of this research is the attributes of the neigh-533borhood environment of the individual house (such as the 534distances to downtown, to schools, to the nearest MRT 535station, to the nearest train station). This means that the 536attributes of the neighborhood environment are variables 537of Level 1. If thinking about taking “urban renewal proj-538ects are nested within various neighborhood areas” for de-539sign, it will set the neighborhood environment as the third 540level of variables. If you are thinking about three levels of 541analysis, it means the individual house is nested in cases 542of urban renewal, and cases of urban renewal are nested 543in the neighborhood. Now we must deal with the problem 544of measuring the neighborhood–how to divide the neigh-545borhood–which was unclear in Taiwan. This time the 546variables that represent the neighborhood environment 547of each neighborhood are the distance of the neighbor-548hood to downtown, to schools, to the nearest MRT sta-549tion, to the nearest train station. If these are the variables 550of the neighborhood environment, it creates the problem 551of overlapping variables of the neighborhood environment 552from Level 1. Analysis on three levels is complicated, and 553for Level 3 we must consider how to represent the char-554acteristic variables of Level 3. This analysis will be more 555complicated than ever. However, if we consider taking 556Level 1 as the individual house, Level 2 as the cases of 557urban renewal, and Level 3 as the administrative area, this 558is more feasible, and we suggest this should also be the 559direction of future research.560Acknowledgements561Thanks to Miss Dai-juan Liu for helping to draw Figure 1.562References563Ahlfeldt, G. M., & Maennig, W. (2010). Impact of sports arenas on 564land values: evidence from Berlin. The Annals of Regional Sci-565ence, 44(2), 205–227. https://doi.org/10.1007/s00168-008-0249-4566Arribas, I., García, F., Guijarro, F., Oliver, J., & Tamošiūnienė, R. 567(2016). Mass appraisal of residential real estate using mul-568tilevel modelling. International Journal of Strategic Property 569Management, 20(1), 77–87. 570https://doi.org/10.3846/1648715X.2015.1134702571Baltagi, B. H., & Bresson, G. (2011). Maximum likelihood es-572timation and Lagrange multiplier tests for panel seemingly 573unrelated regressions with spatial lag and spatial errors: an 574application to hedonic housing prices in Paris. Journal of Ur-575ban Economics, 69(1), 24–42. 576https://doi.org/10.1016/j.jue.2010.08.007577