1# RECORD 281b355bd08bd1b8b228829d9c97b819d0d2a4d12# TITLE: IMPACT OF AIR POLLUTION ON PROPERTY VALUES : A HEDONIC PRICE STUDY3# YEAR: 2013.0 VENUE: nan DOI: 10.23917/jep.v14i1.1504# AUTHORS: Endah Saptutyningsih5# PAGES: 14 CANDIDATE TABLE PAGES (1-based): [9, 10]6# ABSTRACT: The main purpose of this study is the calculation of implicit prices of the environmental level of air quality in Yogyakarta on the basis of housing property prices. By means of Geographical Information System, the housing property prices characterized from the area which have highest air pollution level in province of Yogyakarta. Carbon monoxide is used as the pollution variable. The methodological framework for estimation is based on a hedonic price model. This approach establishes a relationship between the price of a marketable good (e.g. housing) and the amenities and characteristics this good contains. Therefore, if variations in air pollution levels occur, then households would change their behavior in an economic way by offering more money for housing located in highly improved environmental areas. The hedonic regression results that the housing price decrease while increasing the level of air contamination such substance as carbon monoxide.789===== PAGE 1 (score 6, numeric-paren 0, table-words 0, var-words 6) =====1011Jurnal Ekonomi Pembangunan 12Volume 14, Nomor 1, Juni 2013, hlm. 52-65 13 14IMPACT OF AIR POLLUTION ON PROPERTY VALUES : 15A HEDONIC PRICE STUDY 16Endah Saptutyningsih 17Fakultas Ekonomi Universitas Muhammadiyah Yogyakarta 18Jalan Lingkar Selatan, Tamantirto, Kasihan, Bantul, Yogyakarta 19E-mail: end_naufal@yahoo.com 20Diterima 30 Februari 2013 / Disetujui 4 April 2013 21Abstract: The main purpose of this study is the calculation of implicit prices of the environ-22mental level of air quality in Yogyakarta on the basis of housing property prices. By means of 23Geographical Information System, the housing property prices characterized from the area 24which have highest air pollution level in province of Yogyakarta. Carbon monoxide is used as 25the pollution variable. The methodological framework for estimation is based on a hedonic 26price model. This approach establishes a relationship between the price of a marketable good 27(e.g. housing) and the amenities and characteristics this good contains. Therefore, if variations 28in air pollution levels occur, then households would change their behavior in an economic way 29by offering more money for housing located in highly improved environmental areas. The 30hedonic regression results that the housing price decrease while increasing the level of air con-31tamination such substance as carbon monoxide. 32Keywords: air pollution, hedonic price, property value, marginal willingness to pay curve 33JEL: Q53 34Abstrak: Tujuan utama studi ini menghitung harga implisit dari tingkat kualitas lingkungan 35udara di Yogyakarta dengan berbasis pada harga properti perumahan. Dengan menggunakan 36sistem informasi geografis, harga properti perumahan dicirikan dengan daerah yang memiliki 37tingkat polusi udara tertinggi di provinsi Yogyakarta. Karbon monoksida digunakan sebagai 38variabel polusi. Kerangka metodologis untuk estimasi didasarkan pada model harga hedonis. 39Pendekatan ini menetapkan hubungan antara harga barang yang dipasarkan (misalnya 40perumahan) dan fasilitas dan karakteristik barang. Oleh karena itu, jika terjadi variasi dalam 41tingkat polusi udara, maka rumah tangga akan mengubah perilaku mereka secara ekonomi 42dengan menawarkan lebih banyak uang untuk perumahan terletak di daerah lingkungan yang 43baik. Hasil regresi hedonis menunjukkan bahwa penurunan harga perumahan akan menaik-44kan tingkat pencemaran udara seperti misalnya karbon monoksida. 45Kata kunci: polusi udara, harga hedonis, nilai properti, kurva kesediaan marjinal membayar 46JEL: Q53 47 48INTRODUCTION 49Urban pollution is a major focal point of public 50health concern and regularity activity (WHO, 512002). Pertinent guidelines and effective inter-52ferences require knowledge of the burden of 53sickness and premature deaths attributable to 54specified pollutants. A long-term worldwide 55research program is focusing on effects of par-56ticulates and other criteria pollutants such as 57NO2, O3 and CO (Brunekreef and Holgate, 582002). 59Among several tribes of valuation meth-60ods, the hedonic price technique has been 61widely used to assess the effect of urban air 62pollution on property values in developed 63urban centers. Unfortunately, applications ad-64dressed for developing urban centers are rare, 65and very little known about the effect of air 666768===== PAGE 9 (score 44, numeric-paren 20, table-words 7, var-words 10) =====6970Jurnal Ekonomi Pembangunan Volume 14, Nomor 1, Juni 2013: 52-65 7160 72ronmental characteristic CO and the property 73price. All the structural parameters included in 74the model, like plot area and building area, are 75expected to have positive relations with the 76property price. Neighborhood characteristics 77like distance from school and hospital are 78inversely related with property price. It is 79normally expected that as distance from 80hospital and school increases, property price 81decrease, where as distance from city increases 82property price also increases. Applying these 83assumptions on the model specified above, the 84parameters are estimated using the method of 85ordinary least squares and is given in Table 4. 86 87Table 4. Regression Result of Hedonic Price 88Function 89Variable 90Full Model 91Fit Model 92(Constant) 933.323 94(3.509)*** 952.783 96(3.178)*** 97LnLANDSIZE 980.553 99(7.840)*** 1000.528 101(7.658)*** 102LnBUILDSIZE 1030.113 104(1.380) 1050.164 106(2.093)** 107LnNUMROOM 1080.094 109(1.179) 110- 111WSTRUCTDUM 112-0.067 113(-1.088) 114- 115LnDISSCHOOL 116-0.071 117(-1.809)* 118-0.068 119(-1.910)** 120LnDISHPITAL 121-0.085 122(-1.980)** 123-0.075 124(-1.885)* 125LnDISSPMARKET 1260.023 127(0.560) 128- 129 130LnDISREST 131-0.023 132(-0.652) 133- 134 135LnDISCITY 1360.054 137(1.274) 1380.064 139(1.538) 140LnDISMSTREET 1410.024 142 (1.030) 143- 144GARDENDUM 1450.129 146(1.673)* 147- 148LnCO 149-1.305 150(-13.115)*** 151-1.375 152(-15.066)*** 153No. Observation 154Adjusted R2 155250 1560.766 157250 1580.763 159 160The main inferences of this analysis are given 161below: 162(1) All the neighborhood and structural char-163acteristics except building size, wall structure, 164number of rooms, distance from supermarket, 165distance from restaurant, distance from city, 166and distance from main street are statistically 167significant at 5 percent level of significance. 168(2) The environmental characteristic, CO, is 169negatively related to house price and it signifi-170cant at 1 percent level. The results also confirm 171that as the level CO increases by one percent 172house price reduces by 1.375 percent on the 173average. 174(3) Among the neighborhood characteristics, 175the distance from city positively related to 176property price. It showing that when distance 177increases, property price also increases. 178(4) Distance from school or hospital are nega-179tively related to house price, showing that the 180plots nearer the school or hospital have high 181property values. 182(5) The plot area and bulding area are also 183positively related to property price. 184 185The regression results of the hedonic price 186function shows that all the significant estimated 187variables follow the expected relationship pat-188tern. Hence the estimated equation could be 189written as: 190 191 ln PPRICE = 2.783 + 0.528 lnLANDSIZE + 192 1930.164 ln BUILDSIZE – 194 1950.068 ln DISSCHOOL – 196 1970.075 ln DISHPITAL – 198 1991.375 ln CO + e 200 201Calculation of Implicit Marginal Price 202The first derivative of the hedonic price 203function can be interpreted as the implicit 204marginal price function for the environmental 205goods. Descriptive statistics of implicit marginal 206prices for 250 observations is given in the table 2075. 208 209Table 5. Descriptive statistics of Implicit 210 211Marginal Price 212Descriptive Statistics 213Implicit Marginal Price 214(Rupiahs) 215Mean 21620.592,0 217Standard error 2181.133,2 219Median 22015.147,9 221Mode 22213.200,0 223Standard deviation 22417.917,4 225Minimum 2261.555,7 227Maximum 228141.428,5 229 230231232===== PAGE 10 (score 42, numeric-paren 24, table-words 4, var-words 10) =====233234Impact of Air Pollution (Endah Saptutyningsih) 23561236Hence the marginal implicit price for 237reducing CO is calculated as Rp20.600,00. This 238result clearly identifies air quality as an 239important factor, along with structural and 240neighbourhood characteristics, in determining 241demand for property transaction in Daerah 242Istimewa Yogyakarta. 243 244Implicit Demand Function 245As mentioned earlier, second stage estimation 246of inverse demand curve is conducted by 247regressing the implicit marginal price on the 248quantity of environmental goods purchased 249and other socio economic features including 250income level of the individuals. The results are 251given in table 6. 252 253Table 6. Estimation of Inverse Demand Function 254Variable 255Full Model 256Fit Model 257(Constant) 2583.449 259(3.617)*** 2603.022 261(3.425)*** 262INCOMELEVEL 2630.097 264(1.515) 2650.079 266(1.351) 267LnFAMEMBER 268-0.079 269(-0.939) 270- 271LnLANDSIZE 2720.551 273(7.819)*** 2740.533 275(7.727)*** 276LnBUILDSIZE 2770.104 278(1.266) 2790.147 280(1.863)* 281WALLSTRUCTDUM 282-0.062 283(-0.998) 284- 285LnNUMROOM 2860.083 287(1.015) 288- 289LnDISSCHOOL 290-0.065 291(-1.640) 292-0.067 293(-1.889)* 294LnDISHOSPITAL 295-0.085 296(-1.972)** 297-0.080 298(-2.026)** 299LnDISSUPERMARKET3000.024 301(0.584) 302- 303LnDISRESTAURANT 304-0.024 305(-0.685) 306- 307LnDISCITY 3080.056 309(1.337) 3100.071 311(1.722)* 312LnDISMAINSTREET 3130.031 314(1.292) 315- 316GARDENDUM 3170.108 318(1.385) 3190.177 320(1.554) 321LnCO 322-0.317 323(-3.156)*** 324-0.349 325(-3.766)*** 326No. Observation 327Adjusted R2 328250 3290.651 330250 3310.646 332 333The major inferences are: 334(1) The first derivative of the implicit marginal 335price function with respect to CO is negative (-3360.349) signaling decreasing marginal implicit 337prices for increasing environmental quality. It 338means that in the study area, a reduction in CO 339by one percent leads to 0.349 percent increase in 340property values. 341(2) The coefficient of number of family mem-342bers, wall structure, distance from supermarket, 343and 344distance 345from 346main 347street 348are 349not 350significant. 351(3) All other variables except number of family 352members, 353wall 354structure, 355distance 356from 357supermarket, and distance from main street are 358significant at 5 percent level of significance. It is 359interesting to note that the distance from city 360increases by one percent, residential property 361value enhances by 0,071 percent. 362(4) Plot area and building area are positively 363related to residential property values. 364 365So far we have explained the hedonic price 366function estimation and its implications to 367residential property values. The results indicate 368very clearly that the households are willing to 369pay for improved air quality. It is therefore 370necessary to estimate the welfare benefits 371accruing to them trough the purchase of 372property with reduced air quality. 373CONCLUSION 374The primary objective of this study was to 375establish the relationship between air quality 376and the residential property values di province 377of Daerah Istimewa Yogyakarta. By mapping 378the area which has highest concentration of CO 379earlier and estimating the hedonic property 380value in the area, we can establish the relation-381ship between air quality and the residential 382property values. 383In the model specification, we incorporated 384a number of structural, neighborhood, envi-385ronmental and socioeconomic variables as 386determinants of consumer’s willingness to pay 387for reduced air quality. We hypothesized that 388the major environmental variable CO is 389inversely related to the residential property 390values. Similarly, the presence of garden, dis-391tance 392from 393main 394street, 395distance 396from 397398399===== PAGE 11 (score 14, numeric-paren 0, table-words 0, var-words 14) =====400401Jurnal Ekonomi Pembangunan Volume 14, Nomor 1, Juni 2013: 52-65 40262 403supermarket, plot area, and building area are 404positively related to property price while 405distance from hospital and distance from 406restaurant are negatively related. 407By adopting a two-stage estimation proce-408dure to estimate these relationships, we found 409that, on the average, an increase in the level of 410CO reduced property prices in the study area 411by 1.375 percent. We estimated the marginal 412implicit price for reducing CO as Rp 20.600,00. 413Estimates further revealed that the households 414are willing to pay an additional amount of 0.35 415percent for a reduction in CO. In short, the 416analysis revealed a positive response of house-417holds in Daerah Istimewa Yogyakarta -which 418have highest concentration of CO - between air 419quality and property prices. 420REFERENCES 421Bajari, P and CL Benkard. 2005. “Demand Esti-422mation with Heterogeneous Consumers 423and Unobserved Product Characteristics: 424A Hedonic Approach”. Journal of Political 425Economy, 113(6), 1239-1276. 426Bhattacharya, Rabindra, N. 2002. Environmen-427tal Economics: An Indian Perspective. New 428Delhi: Oxford University Press. 429Boyle, M. and K. Kiel. 2001. “A Survey of House 430Price Hedonic Studies of the Impact of 431Environmental Externalities”. Journal of 432Real Estate Literature. Vol. 9, No 2, pp. 117-433144. 434Briggs, D.J.; de Hoogh, C.; Gulliver, J.; Wills, J.; 435Elliott, P.; Kingham, S.; Smallbone, K. 4362000. “A Regression based Method for 437Mapping Traffic-related Air Pollution: 438Application and Testing in Four Con-439trasting Urban Environments.” Sci Total 440Environ 253(1-3): 151-167. 441Cicero-Fernandez, P.; Torres, V.; Rosales, A.; 442Cesar, H.; Dorland, K.; Munoz, R.; Uribe, 443R.; Martinez, A.P. 2001. “Evaluation of 444Human Exposure to Ambient PM10 in the 445Metropolitan area of Mexico City using a 446GISbased Methodology”. J Air Waste 447Manag Assoc 51: 1586-1593. 448Civis, S, Z. Zelinger, M. Strizik, Z. Janour. 2001. 449“Simulation of Air Pollution in a Wind 450Tunnel”. In: Spectroscopy from Space. 451(Demaison, J., Ed.), Kluwer Academic, 452Dordrecht, 275-299. 453Cowell PJ and Zeng TQ. 2003. “Integrating Un-454certainty Theories with GIS for Modeling 455Coastal Hazards of Climate Change.” 456Marine Geodesy, 26, 5-18. 457Gualtieri G., & Tartaglia M. 1998. Predicting 458Urban Traffic Air Pollution: a GIS frame-459work, Transportation Research - D, 3(5), pp. 460329-336, Elsevier Science. 461Hoek, G.; Fischer, P.; Van Den Brandt, P.; 462Goldbohm, S.; Brunekreef, B. 2001. “Esti-463mation of Long-term Average Exposure 464to Outdoor Air Pollution for a cohort 465study on mortality.” J Expo Anal Environ 466Epidemiol 11(6): 459-469. 467Holgate. 2002. Air Pollution and Health. Lancet, 468360: 1233-1242. 469Janour, Z., Z. Zelinger, S. Civis. 1999. “Laser 470Photoacoustic Spectrometry and its Ap-471plication for Simulation of Air Pollution 472in a wind tunnel”. Analyst, 124, 1205-1208. 473Jones C B, Kidner D B, Luo L Q, Bundy Gl And 474Ware J M.. 1996. Database design for a 475Multi-scale Spatial Information System. 476Int. Journal of GIS, 10(8), pp.901-920. 477Lave, L.B. and Seskin, E.P., 1972. “Air pollution, 478climate, and home heating: their effects on 479US mortality rates”. Am J Public Health 62 4807, pp. 909–916 481Longley, P.A., M.F. Goodchild, D.J. Maguire, 482D.W. Rhind. 2001. “Geographic Infor-483mation Systems and Science.” John Wiley 484& Sons, 27- 58, New York. 485Moaz, Alsherfawi Aljazaerli. 2005. “Hedonic 486Valuation of Marginal Willingness to Pay 487for Air Quality in Metropolitan Damas-488cus.” Forum of International Development 489Studies, 3 September 2005. 490Murty, M., S. Gulati and A. Banerjee. 2003. 491Hedonic Property Prices and Valuation of 492Benefits from Reducin Urban Air Pollu-493