1# RECORD e7545aa69fd7e3b0b4fddbe8806b08519280a6742# TITLE: Analysis of House Prices to Assess Economic Impacts of New Public Transport Infrastructure3# YEAR: 2011.0 VENUE: nan DOI: 10.3141/2245-164# AUTHORS: Lucía Mejía Dorantes; A. Páez; J. M. Vassallo5# PAGES: 9 CANDIDATE TABLE PAGES (1-based): [7]6# ABSTRACT: Transportation infrastructure is known to affect the value of real estate property by virtue of changes in accessibility. The impact of transportation facilities is highly localized as well, and it is possible that spillover effects result from the capitalization of accessibility. The objective of this study was to review the theoretical background related to spatial hedonic models and the opportunities that they provided to evaluate the effect of new transportation infrastructure. An empirical case study is presented: the Madrid Metro Line 12, known as Metrosur, in the region of Madrid, Spain. The effect of proximity to metro stations on housing prices was evaluated. The analysis took into account a host of variables, including structure, location, and neighborhood and made use of three modeling approaches: linear regression estimation with ordinary least squares, spatial error, and spatial lag. The results indicated that better accessibility to Metrosur stations had a positive impact on real estate values and that the effect was marked in cases in which a house was for sale. The results also showed the presence of submarkets, which were well defined by geographic boundaries, and transport fares, which implied that the economic benefits differed across municipalities.789===== PAGE 1 (score 13, numeric-paren 0, table-words 0, var-words 13) =====1011Analysis of House Prices to Assess 12Economic Impacts of New Public 13Transport Infrastructure 14Madrid Metro Line 12 15Lucia Mejia Dorantes, Antonio Paez, and José Manuel Vassallo 16Transportation infrastructure is known to affect the valué of real estáte 17property by virtue of changes in accessibility. The hnpact of transporta-18tion facilities is highly localized as well, and it is possible that spillover 19effects result from the capitalization of accessibility. The objective of this 20study was to review the theoretical background related to spatial hedonic 21models and the opportunities that they provided to evalúate the effect of 22new transportation infrastructure. An empirical case study is presented: 23the Madrid Metro Line 12, known as Metrosur, in the región of Madrid, 24Spain. The effect of proximity to metro stations on housing prices was 25evaluated. The analysis took into account a host of variables, including 26structure, location, and neighborhood and made use of three modeling 27approaches: linear regression estimation with ordinary least squares, 28spatial error, and spatial lag. The results indicated that better accessibility 29to Metrosur stations had a positive hnpact on real estáte valúes and that 30the effect was marked in cases in which a house was for sale. The results 31also showed the presence of submarkets, which were well defined by 32geographic boundaries, and transport fares, which implied that the 33economic beneflts differed across municipalities. 34Public transportation infrastructure is known to affect urban áreas in 35a number of significant way s. The effects of public transportation were 36classifiedby Banister andBerechman (i) as follows: transportation-37related effects, land use effects, and effects on jobs and business 38activities. Researchers, including Mas andMaudos, have shownthat 39transportation infrastructure has significant, direct effects that also tend 40to spill over in the form of neighborhood effects (2). According to 41Boarnet (3), these effects can be classified as short term (i.e., related to 42reductions in travel time for the population that lives or works around 43the infrastructure), and long term [i.e., that stem from agglomera-44tion economies caused by positive externalities that improve the 45efficiency of firms as the result of economies of scale (4)]. 46Commuter rail and metro stations can have both positive 47(e.g., accessibility benefits) and negative (e.g., noise and congestión) 48L. M. Dorantes and J. M. Vassallo, Departamento de Transportes, Universidad 49Politécnica de Madrid, ETSI de Caminos, Canales y Puertos, Profesor Aranguren s/n. 5028040 Madrid, Spain. A. Paez, School of Geography and Earth Sciences, McMaster 51University, General Science Building, Room236, 1280 Main Street West, Hamilton. 52Ontario L8S 4L8, Canadá. Corresponding author: L. M. Dorantes, lmejia@caminos. 53upm.es 54effects. On the positive side, metro stations increase accessibility to 55public transportation for the people that live nearby and reduce their 56travel time and costs to other destinations in urban áreas, which 57produces a network externality. Similarly, business activities near 58the stations enjoy advantages. Jobs and shops are accessible to those 59that come from any destination. Business activities cióse to the 60stations—especially shops—benefit from the increase in passersby 61to and from the stations. Many of the negative effects associated 62with metro stations are spatial externalities, which include noise and 63changes to the urban landscape. The effects described abo ve depend on 64the type of transport infrastructure (2, 5), its location, and its specific 65characteristics. Researchers such as Maoh et al. have noted a strong 66interdependence between the locational patterns among residential 67and commercial land development (6). 68A problem of interest that arises whenever new transportation 69infrastructure, services, or both, are introduced is how to assess the net 70impact of the positive and negative effects. Two methods to assess 71real estáte markets are discussed in the literature: the repeated sales 72approach and the hedonic price analysis. Hybrid models, which 73combine those methodologies, also have been used. Bailey et al. first 74presented repeated sales deals with a regression of the difference 75in sale prices for the same set of homes and a set of time-of-sale 76indicators (7). Repeat sales can be difficult to implement as the result 77of small sample sizes. The sale of each property may be transacted 78infrequently, and the lengthy periods of time needed to collect a 79sufficiently large sample can be impractical or impossible in the 80case of new infrastructure. Compared with hedonic price analysis, 81repeat sales implementations are relatively rare (8). 82The hedonic models refer to the implicit characteristics of a good 83(9). In the case of real estáte, the idea is that property prices can be 84decomposed to evalúate the price of various structural, locational, 85and neighborhood characteristics. To estimate the implicit price for 86each attribute, the price is modeled as a function of various attributes 87and the use of a linear regression model with parameters estimated 88by the ordinary least squares (OLS) method. A house unit is fixed in 89its location and influenced by the characteristics of the neighborhood, 90which constitute a set of geographically distributed data. Data 91members may be related to each other as well, which is known as 92spatial autocorrelation (10,11). Spatial hedonic analysis has emerged 93as an approach that accounts for spatial autocorrelation in real estáte 94applications. 95The literature that accounts for transport effects in land valué 96through different approaches is extensive (12-16). As noted by 97Martínez and Viegas (17) andDu andMulley (18), however, studies 9899100===== PAGE 7 (score 62, numeric-paren 51, table-words 2, var-words 7) =====101102TABLE 2 103Hedonic Model Results 104if-squared 105Log likelihood 106Akaike info 107criterion 108Variable 109C 110m2 111lift 112bathr 113heating 114a a 115basem 116spool 117parking 118zone B2 119ms/pci 120cer/pci 121hosp/pci 122pop 123lambda 124W_price 125.7591 126-20,249.9 12740,527.8 128OLS 12965,921.5600** 130(3,975.429) 1311,402.8910** 132(42.63756) 13319,696.7800** 134(2,053.673) 13517,679.4900** 136(2,281.784) 1375,868.9630** 138(1,855.498) 1395,675.9520** 140(1,628.007) 1418,831.1460** 142(1,909.752) 14315,517.4800** 144(2,313.651) 14524,322.4000** 146(2,220.907) 147-31,553.56** 148(1,899.969) 149-45,862.7400* 150(18,787.84) 151-76,497.57** 152(13,989.91) 15361,411.55** 154(18,107.78) 155-0.1244867** 156(0.04346661) 157.7636 158-20,238.6 15940,505.3 160Spatial Error Model 16168,564.3** 162(4,265.663) 1631,400.5710** 164(43.10532) 16519,308.1700** 166(2,083.496) 16716,844.8700** 168(2,293.362) 1695,214.945** 170(1,889.629) 1715,603.8660** 172(1,601.317) 1738,139.1700** 174(1,907.114) 17515,329.5700** 176(2,406.694) 17723,780.6500** 178(2,252.612) 179-31,564.53** 180(2,246.84) 181-45,912.3700* 182(22,044.94) 183-74,584.5** 184(16,442.75) 18560,944.1** 186(21,288.32) 187-0.1474505** 188(0.05079898) 1890.1827401** 190(0.0379598) 191.7650 192-20,231.4 19340,492.8 194Spatial Lag Model 19542,584.1** 196(5,484.156) 1971,357.3570** 198(42.76402) 19918,207.8600** 200(2,035.105) 20115,889.0300** 202(2,260.322) 2035,113.653** 204(1,827.406) 2055,568.1830** 206(1,601.539) 2078,132.9380** 208(1,879.831) 20912,687.4200** 210(2,324.918) 21122,977.9400** 212(2,194.465) 213-28,817.85** 214(1,924.455) 215-42,145.6000* 216(18,492.35) 217-69,047.84** 218(13,819.2) 21947,469.47** 220(17,963.27) 221-0.07174224 222(0.04380104) 2230.1338048** 224(0.02199083) 225NOTE: Estimated coefficients are presented in euros; observations = 1,714. 226*p<.05;**p<.0l. 227€140,000.00 228€135,000.00 229€130,000.00 230€125,000.00 231€120,000.00 232€115,000.00 233€110,000.00 234€105,000.00 235€100,000.00 236-ALCORCON 237-GETAFE 238-LEGANES 239-MOSTOLES 240-FUENLABRADA 2411000 242Distan ce (m) 243(a) 244€150,000.00 245€140,000.00 246€130,000.00 247€120,000.00 248€110,000.00 249€100,000.00 250€90,000.00 251€80,000.00 252-LEGANES 253-ALCORCON 254-GETAFE 255-MOSTOLES 256-FUENLABRADA 257500 2581000 259Distance (m) 260( ) 261FIGURE 4 262Impact on valué of house related to its distance from (a) Metrosur and Ib) Cercanías stations. 263264265===== PAGE 8 (score 18, numeric-paren 0, table-words 0, var-words 18) =====266267and circular accessibility. Thus transfer stations appeared to have 268become valued the most, ceteris paribus, among all stations within 269the municipalities studied. 270CONCLUDING REMARKS AND DISCUSSION 271This paper presents an empirical analysis of the real estáte market 272in the south of Madrid to assess the economic benefits of a recent 273metro line infrastructure. The results of the study indicated that 274better accessibility to Metrosur stations had a positive impact on real 275estáte valúes and that the effect was particularly marked in cases in 276which a house was for sale. 277To avoid a discussion of pecuniary impact (46, 47), the results of 278the three models (OLS, SLM, and SEM), are presented, which were 279similar. According to Tsutsumi and Seya, some limitations exist in 280the spatial econometric approach to benefit evaluation because it uses 281a spatial weight matrix. Thus those parameters would be valid only 282for the given spatial weight matrix (54). In an empirical analysis, 283however, it is possible to observe that results in all cases are quite 284similar. It was concluded that the estimates were similar because the 285models accounted for housing submarkets and use was made of a 286detailed data set. This conclusión was consistent with Bourassa 287et al. (44, 45). Even without a spatial analysis, if submarkets are well 288controlled (in this case, market segmentations were accounted for 289by means of geographic boundaries and the transit zonal fare), and 290accurate information is used, OLS is a practical method that pro vides 291congruent estimates. 292The results successfully appraised the benefit that an urban mass 293transport infrastructure produces in house prices. The distance to 294Cercanias stations, however, was economically more important that 295the distance to the Metrosur stations. This result may be explained 296by the fact that the commuter rail fines took people directly to down-297town Madrid through a radial trip and therefore in less time. Metrosur 298stations took people to other municipalities or to the closest metro 299and commuter transfer station. Apparently, those commuters that 300used public transportation preferred transit that took them to Madrid 301(i.e., Cercanias, commuter rail fines). 302The municipalities that benefited the least from Metrosur had a more 303expensive transit fare (Zone B2), were located farthest away from the 304city of Madrid, and had the smallest per capita income. These obser-305vations revealed the trade-off between lower prices for housing and 306higher costs in time and transport fares. In other words, the results 307showed a higher willingness to pay for more accessible locations, 308especially in the municipalities closest to Madrid. 309Sometimes fare policies are intended to limit the location of people 310far from the city core. From the perspective of social equity, however, 311continued research might be needed to verify whether this policy 312fulfills the transportation goals at the local and regional levéis, or 313whether it limits the integration of outlying populations with the 314whole región. Researchers such as Cervero have found that transit 315fares have different effects in áreas with low population density and 316suburban settings: people are more reluctant to take public transport 317if they have a car available and at the same time are prone to accept 318an increase in the ticket fare when they make radial trips, which in this 319case would mean a higher appraisal of Cercanias than Metrosur (55). 320The results with regard to parking spaces also suggested that people in 321this área relied on cars to commute, which was in line with the findings 322of Jordá (24). The appraised valué of transfer stations increased in 323both modes, which related to an increase of vertical and horizontal 324accessibility. Further studies should focus on the advantages that 325populations perceive in these stations. 326If the reason to travel from one point to another many times is to 327engage in economic activities, such as shopping and working, the 328question that arises is if the increase in economic activities within 329the five municipalities is sufficient to cause people to increase their 330use of Metrosur around the región rather than go back and forth 331to the city of Madrid. A better job-housing balance would increase the 332use of Metrosur and therefore its economic appreciation. Rather 333than foster integration across all the municipalities, Fuenlabrada and 334Mosto les might pursue a sepárate, less accelerated track in the quest 335for more balanced growth. 336Urban planning and design play important roles to foster the use of 337public transport. A dense street network that puts public transit within 338easy reach encourages its use. A look at new urban development in 339the zones that were studied, however, showed that they were car-340oriented. Examples, such as the A-B-C policy in the Netherlands, 341should be borne in mind, which have succeeded in taking advantage 342of áreas cióse to transit stations. Not only is it important to have a 343public transit station near a neighborhood but also a combination 344of other factors in place, such as the optimization of resources and 345political agreements that foster the use of public transport. 346Finally, another application of these models for urban studies might 347be valué capture. Although the subject of much discussion, it is a com-348plicated concept to put it into practice. In Spain, certain fiscal mecha-349nisms already are in place that relate to property prices (56). The Urban 350Land Valué Increase Tax (commonly known as Impuesto sobre Plus-351valías) is one of them. Its main problem is that Spanish taxes are not 352earmarked. None of the benefits triggered by the transport infrastruc-353ture are returned to the Madrid regional transport authority, because it 354is the municipality that levies the rate of the tax and is the one that col-355lects it. These kinds of models open the possibility to account for real 356increases in land valué and any transport impaets in order to transfer to 357the transport authority the percentage increase produced by the public 358transportation sy stem. Such an approach would lead to more economic 359resources to maintain the public transport infrastructure with higher 360social acceptability, especially during periods of budgetary constraints. 361ACKNOWLEDGMENTS 362Financial support from the Ministry of Science and Innovation of 363Spain is gratefully acknowledged. The authors thank Javier Gutiérrez 364and Juan Carlos Garcia of the Department of Human Geography of 365the Universidad Complutense de Madrid for information that helped 366to carry out this research. 367REFERENCES 3681. Banister, D., and J. Berechman. Transport Investment and Economic 369Development. Routledge, New York, 2000. 3702. Mas, M., and J. Maudos. Infraestructuras y Crecimiento Regional En 371España Diez Años Después. In Competitividad Regional en la Union 372Europea Ampliada (José Villaverde Castro, coord.), Instituto de Estudios 373Fiscales, Madrid, 2004, pp. 143-168. 3743. Boarnet, M. G. Conducting Impact Evaluations in Urban Transport. 375Doing Impact Evaluation Series 5. World Bank, Washington, D.C., 2006. 3764. Fujita, M. Urban Economic Theory: land Use and City Size. Cambridge 377University Press, United Kingdom, 1989. 3785. Cambridge Systematics Inc. TCRP Report 35: Economic Impact Analysis 379of Transit Investment: Guidebook for Practitioners, TRB, National 380Research Council, Washington, D.C, 1998. 3816. Maoh, H. F., M. Koronios, and P. Kanaroglou. Exploring the Land 382Development Process and Its Impact on Urban Form in Hamilton, Ontario. 383The Canadian Geographer, Vol. 54, No. 1, 2010, pp. 68-86. 3847. Bailey, M. J., R. F. Muth, and H. O. Nourse. A Regression Method for 385Real Estáte Price Index Construction. Journal of the American Statistical 386Association, Vol. 58, No. 304, 1963, pp. 933-942. 387