1# RECORD 84a6e8ef31bf481e52cb738679f636c9a2fcc69f2# TITLE: Loss of Value in House Properties due to Air Pollution in Map Ta Phut Industrial Area of Thailand3# YEAR: 2016.0 VENUE: nan DOI: 10.35762/aer.2016.38.3.14# AUTHORS: Nararuk Boonyanam5# PAGES: 13 CANDIDATE TABLE PAGES (1-based): [8, 9, 10]6# ABSTRACT: Air pollution in the Map Ta Phut Industrial Estate in Thailand, has caused increasing risk of death, injury and health problems for local residents. Existing control measures are ineffective due to weak law enforcement. Although the problem has drawn a lot of attention, the monetary value of the impacts of air pollution in Map Ta Phut is not known. The objective of this paper was to estimate the economic cost of air pollution damage in Map Ta Phut using the Hedonic Pricing Method, with 192 house samples. The results showed that the marginal willingness to pay (MWTP) for a one unit reduction of AQI level was THB 2,815 or around USD 80, or each unit increase in air pollution would result in an estimated loss of THB 2,815 or around USD 80 in house price. The total value of losses to housing from one unit of air pollution in Map Ta Phut which is comprise of 38 communities was THB 119.3 million (about USD 3.4 million). Detection of the loss of value has implications for local residents, housing development investors, factories, and government and environmental institutions to manage air pollution as their top priority as well as to evaluate the efficiency of existing measures and budget to deal with the problem more appropriately.789===== PAGE 1 (score 5, numeric-paren 0, table-words 0, var-words 5) =====1011App. Envi. Res. 38 (3): 1-13 12 13Loss of Value in House Properties due to Air Pollution 14in Map Ta Phut Industrial Area of Thailand 15 16Nararuk Boonyanam 17 18Faculty of Economics at Sri-racha Kasetsart University, Sri-racha, Chonburi 20230, Thailand 19* Corresponding author: Email: noaratc@gmail.com; Phone: +66895312079 20 21Article History 22Submitted: 8 June 2016/ Accepted: 2 August 2016/ Published online: 17 October 2016 23 24 25Abstract 26 27Air pollution in the Map Ta Phut Industrial Estate in Thailand, has caused increasing risk 28of death, injury and health problems for local residents. Existing control measures are inef-29fective due to weak law enforcement. Although the problem has drawn a lot of attention, the 30monetary value of the impacts of air pollution in Map Ta Phut is not known. The objective of 31this paper was to estimate the economic cost of air pollution damage in Map Ta Phut using the 32Hedonic Pricing Method, with 192 house samples. The results showed that the marginal willing-33ness to pay (MWTP) for a one unit reduction of AQI level was THB 2,815 or around USD 3480, or each unit increase in air pollution would result in an estimated loss of THB 2,815 or around 35USD 80 in house price. The total value of losses to housing from one unit of air pollution in Map 36Ta Phut which is comprise of 38 communities was THB 119.3 million (about USD 3.4 million). 37Detection of the loss of value has implications for local residents, housing development in-38vestors, factories, and government and environmental institutions to manage air pollution as 39their top priority as well as to evaluate the efficiency of existing measures and budget to deal 40with the problem more appropriately. 41 42Keywords: Hedonic pricing method; Home price, Air pollution; Map Ta Phut 43 44 45Introduction 46 47Map Ta Phut residents have suffered from 48various forms of pollution, especially air pollu-49tion caused by frequent chemical leaks and emis-50sions from factories in the Map Ta Phut Indus-51trial Estate, located in Rayong province, Thailand. 52Operations began in 1990, and the estate now 53hosts 140 industrial plants, including 45 petro-54chemical factories, 8 coal-fired power plants, 12 55chemical fertilizer factories and 2 oil refineries 56[10]. A total of 56,591 residents (excluding a 57large non-registered population of 102,362), live 58 59Applied Environmental Research 60 61Journal homepage : http://www.tci-thaijo.org/index.php/aer 62 636465===== PAGE 8 (score 23, numeric-paren 0, table-words 4, var-words 11) =====66678 App. Envi. Res. 38 (3): 1-13 68Results 69Descriptive statistics are presented in Table 4, 70which shows that the average house price in Map 71Ta Phut was THB 1,750,461 where the average 72house age was 13 years, and contained an ave-73rage three bedrooms and two bathrooms. Ave-74rage land size was 191.63 m2. Average distance 75to the district center was around 7 km. In addi-76tion, less than half of the sample houses were 77located on the main road. Half of the sample 78houses were detached, and the other half was 79townhouses which provided a representative 80sample of each house type. All these attributes 81showed that the houses in the Map Ta Phut 82area belonged to middle to upper-middle class 83families.84 85Table 4 Descriptive statistics 86Variable 87 88Unit 89Mean 90TD.DEV. 91Dependent variable 92 93PH 94House price 95THB 961,750,461.00 867,495.30 97Structural Characteristics 98 99Detach 100House type 1011=house, 0=townhouse 102 0.49 103 0.50 104Age 105House age 106Year 107 12.84 108 14.02 109Bed 110Number of bedroom 111rooms 112 2.47 113 0.62 114Bath 115Number of bathroom 116rooms 117 1.84 118 0.57 119Area 120Utilization area 121square meter 122 141.49 123 146.64 124Lot 125Land size 126square meter 127 191.63 128 219.12 129Neighborhood Characteristics 130 131Discbd 132Distance to district center 133meter 134 6,740.13 135 7,978.17 136Main 137Main road location 1381=main road, 1390=otherwise 140 0.29 141 0.46 142Air Pollution 143 144SO2 145Hourly average concentration of SO2 146ppb 147 5.05 148 3.91 149NO2 150Hourly average concentration of NO2 151ppb 152 19.29 153 2.93 154CO 1558 hour average concentration of CO 156ppm 157 0.99 158 0.10 159O3 160Hourly average concentration of O3 161ppb 162 32.14 163 1.47 164PM10 165Daily average concentration of PM10 166ug/m3 167 82.57 168 32.76 169VC 170Daily average concentration of VC 171ug/m3 172 0.21 173 0.04 174BT 175Daily average concentration of BT 176ug/m3 177 1.20 178 0.77 179DM 180Daily average concentration of DM 181ug/m3 182 0.79 183 0.01 184CF 185Daily average concentration of CF 186ug/m3 187 0.19 188 0.01 189DE 190Daily average concentration of DE 191ug/m3 192 0.56 193 0.10 194BZ 195Daily average concentration of BZ 196ug/m3 197 2.15 198 0.20 199AQI 200Daily average concentration of 201AQI 202- 203 43.30 204 13.20 205Note: 1 The air pollution level of AQI between 0-50 = Excellent, 51-100 = good, 101-150 = lightly polluted, 206151-200 = moderately polluted, 201-300 = heavily polluted and 300+ = severely polluted 207 208A correlation test showed perfect correlation 209between the air pollution criteria variables as well 210as VOCs substances, therefore only AQI was used 211to represent the ambient level of air pollution and 212the remaining air pollution criteria along with 213the VOCs were therefore omitted from the ana-214lysis to avoid multicollinearity problems that may 215reduce the precision of the estimate. The non-nor-216mality problem was corrected using a boxplot to 217identify data outliers; eight outliers were removed 218from the analysis. After removal of these out-219liers, the total sample was 192. The results of Box-220cox transformations are presented in Table 5.221222223===== PAGE 9 (score 87, numeric-paren 72, table-words 1, var-words 2) =====224225App. Envi. Res. 38 (3): 1-13 9 226Table 5 Hypothesis test for Box-Cox transformation 227Characteristic 228Variables 229Box Cox Transformation 230lhsonly Model 231 rhsonly Model 232Lambda Model 233Theta Model 234Structural Characteristics 235Detach 236434.886*** 237 391,533.500** 2383.048** 23910.738* 240(26.884) 241(5.373) 242(5.310) 243(3.356) 244Age 245-0.398 246-45,908.140 247-0.247 248-2.656 249(0.033) 250(0.300) 251(0.751) 252(0.265) 253Bed 254156.273*** 255286,382.10 2563.289** 25717.640 258(6.852) 259(0.976) 260(4.344) 261(2.115) 262Bath 263262.351*** 264569,925.700*** 2654.093*** 26622.805*** 267(17.368) 268(7.381) 269(10.071) 270(7.779) 271Area 272-0.408* 273-166,003.900 274-0.178 2750.131 276(3.087) 2770.427 278(0.302) 279(0.000) 280Lot 2810.440*** 2821,561,252.000*** 2831.480*** 284241.498*** 285(7.316) 286(9.930) 287(14.003) 288(19.346) 289Neighborhood Characteristics 290Discbd 291-0.029 292-80,836.090 293-0.035 294-13.424 295(1.765) 296(0.106) 297(0.027) 298(0.080) 299Main 300-73.568 301-184,539.000* 302-1.488* 303-6.970 304(1.094) 305(3.260) 306(2.703) 307(2.265) 308Air Pollution 309AQI 3109.955 311-1,144,197.000 312-2.144 313-144.509* 314(0.590) 315(1.101) 316(0.511) 317(3.582) 318Constant 319596.9216 3202488.723 32158.813 3220.810 323AIC 3242180.956 3252187.270 3262175.152 3272170.988 328BIC 3292183.286 3302189.601 3312177.483 3322175.649 333Lambda 334- 335-0.178 3360.175 337-0.461* 338s.e. 339- 340(0.295) 341(0.172) 342(0.324) 343Theta 3440.476*** 345- 3460.175 3470.288** 348s.e 349(0.171) 350- 351(0.172) 352(0.160) 353Loglikelihood 354-1089.478 355-1092.6351 356-1086.5762 357-1083.494 358LR chi-square 359103.740 360107.750 361109.54 362115.710 363Chi-square for reject Ho 364when X = -1 36565.790*** 3664.750** 36751.670*** 36857.830*** 369when X = 0 3707.680*** 3710.350 3721.130 3737.290*** 374when X = 1 3759.120*** 3762.810* 37714.930*** 37821.090*** 379 380 381 382The Akaike information criterion (AIC) and 383the Bayesian information criterion (BIC) were 384used to select the most suitable Box-Cox model 385from the four functional forms. The theta model 386had the lowest AIC and BIC value, suggesting 387that it provides the best transformation. The 388theta model produced a value of θ of 0.288 and 389the value of λ was -0.461 and yielded the best 390transformation. The rejection of parameters test 391of the theta model under the null of θ = λ = 1, 0, 392or -1 indicated that linear and log-log linear mo-393del were not suitable functional forms. In addi-394tion, the lhsononly model test result also sug-395gested that the semi-log functional form should be 396rejected although the model was not of the best 397fit. Since all simpler functional forms were rejected, 398a Box-Cox transformation was used as a func-399tional form for this study. The use of the Box-400Cox transformation as a functional form (not 401just as a test) of choice in hedonic regression has 402403404===== PAGE 10 (score 21, numeric-paren 2, table-words 3, var-words 10) =====40540610 App. Envi. Res. 38 (3): 1-13 407generally received strong support in the litera-408ture [25-26]. The final regression model was as 409shown in Eq.10. 410The regression result showed that four out of 411nine variables were significant, namely: house 412type, number of bathrooms, air quality and land 413size. The signs of the significant variables were 414as expected. The air pollution level was nega-415tively correlated with house price, while house 416type, number of bathrooms and land size were 417positively correlated, confirming both our re-418search hypotheses. The negative sign of the air 419pollution variable indicated that people are 420concerned about air pollution levels. The impli-421cit price for significant continuous characteris-422tics and non-transformed or dummy variables 423can be calculated using equations (8) and (9). 424 425 426 y(0.288) = 0.810 + 10.738Detach + 22.805Bath-0.461 + 241.498Area-0.461 – 144.509AQI-0.461 (Eq. 10) 427 428where y(0.288) = y0.288 – 1, xi-0.461 = x-0.461i – 1, xi = Detach, Bath, Area and AQI 429 0.288 430 -0.461 431 432The implicit price in Table 6 indicates that 433the greatest effect on house price is house type. 434The price difference between detached houses 435and townhouses was around THB 300,000. The 436second highest effect was from number of bath-437rooms, followed by air pollution. It was surpri-438sing to find that the effect of air pollution was 439greater than that of the land size variable. (Air 440pollution factors are typically found to have less 441impact on house price compared to physical com-442ponents and location of the house). The coeffi-443cient of the AQI variable suggested that better 444air quality is associated with a higher house 445price. The result is straightforward. The margi-446nal effect of a change in one unit of AQI or it 447can be interpreted as the marginal willingness to 448pay (MWTP) for a one unit level reduction in 449AQI or air pollution of THB 2,815 or around 450USD 80. The land size variable had the smallest 451effect on house price. One additional square me-452ter of land size will increase the house price by 453THB 2,221. The result is consistent with the land 454price appraisal report of the Treasury Depart-455ment which indicated that the land price per 4564 m2 in the Mapchalood-Charkklang community 457which has a high polluted air quality index was 458THB 10,000, while the land price per 4 m2 in the 459Sanam Pao and Kangamphur-Tangpai commu-460nities which have a low polluted air quality in-461dex was between THB 40,000 and THB 60,000. 462The model can be used to calculate average house 463prices that have certain structural and environ-464mental characteristics which benefit both local 465residents and housing development agents. 466Table 6 Implicit price of significant variables 467Significant 468variable 469Variable type 470Coefficient 471(βi, γi) 472xiλ-1 473y 1-θ 474Implicit price475Detach 476Dummy 47710.738478 47927869.2022 480299,259.4933481Bath 482Continuous 48322.8054840.01037348527869.2022 4866,592.8421487Lot 488Continuous 489241.4984900.00033049127869.2022 4922,221.2702493AQI 494Continuous 495-144.5094960.00069949727869.2022 498-2,814.5760499 500Discussion 501The study results confirm an association be-502tween air pollution and house price in Map Ta 503Phut. The MWTP for a one unit reduction of AQI 504level was calculated as THB 2,815 or around 505USD 80. The result was similar to that found by 506[24] for Jakarta city in Indonesia which reported 507that the MWTP for better air quality was in the 508509510===== PAGE 11 (score 17, numeric-paren 0, table-words 0, var-words 17) =====511512App. Envi. Res. 38 (3): 1-13 11 513range USD 28.34-USD 85.39 or around THB 514995 to THB 2,998. The result obtained from 515Map Ta Phut was quite low in the range com-516pared with evidence from a developed country 517such as the USA (reported by Boyle and Kiel [4] 518as in the SO2 range of USD 58 to USD 328/ 519μg/m3). There are 38 communities in the Map 520Ta Phut Municipality, comprising 42,392 house-521holds. Therefore, the total loss of value in house 522property from air pollution damage in the Map 523Ta Phut Municipality per unit of AQI is around 524THB 119.3 million or about USD 3.4 million. 525This figure is in the housing sector only and is a 526result of one unit change or a marginal change in 527the air pollution damage level. Therefore, in rea-528lity, the actual monetary loss should be much 529higher, as normally the air pollution level changes 530by more than the marginal value. The huge eco-531nomic loss identified by this study indicates that 532the problem should be treated as a top priority 533over public health and safety to avoid tremen-534dous accumulation of economic loss in the fu-535ture. Better budget allocation, better law enforce-536ment and better cooperation should be adopted 537immediately before the crisis grows beyond our 538capability to deal with it. 539 540Conclusions and implications 541House property values in Map Ta Phut de-542pend on their structural characteristics and air 543pollution levels. The structural characteristics 544that affect house price are house type, number of 545bathrooms, air quality and land size. Only an 546aggregate parameter in the form of Air Quality 547Index (AQI) was tested due to the multicolli-548nearity problem between air pollution criteria 549variables as well as VOCs. The type of house 550contributes the highest effect to house price com-551pared to the effect of additional bathrooms, air 552quality and lot size, indicating that people in the 553study area were more concerned about house 554type than number of bathrooms, air quality and 555lot size. Factors such as utility area, house age, 556number of bedrooms, distance to district center 557and location of the house were found to have no 558effect on house price. Housing development agents 559should take the variables with a large effect on 560house price into account–house type, number of 561bathrooms, air quality level and lot size. –when 562developing a housing project. Local residents 563should consider relocating or selling their house 564in Map Ta Phut area to prevent further loss to 565their house value if air pollution levels show no 566sign of improvement. By recognizing the large 567loss from air pollution damage in Map Ta Phut, 568the government, environmental institutions and 569industrial factories should be able to better ma-570nage their priorities and be more concerned about 571dealing with the air pollution level more appro-572priately. 573 574References 575[1] Ara, S., Tekes, C. 2016. The monetary va-576luation of acute respiratory illness from air 577pollution in Turkey. Atmospheric Pollu-578tion Research, 7(2016): 82–91. 579[2] 580Aungudompukdee, P. 2009. Factors re-581lated to neurobehavioral effects in young 582children residing near petrochemical in-583dustrial estate: A community-based cross-584sectional study in Map Ta Phut sub-dis-585trict, Rayong province, Thailand. Ph.D. 586dissertation, Chulalongkorn University. 587[3] 588Boonpraderm, C. 2010. The implementa-589tion effectiveness of the air pollution con-590trol policy of the Map Ta Phut Industrial 591Estate. Ph.D. dissertation, Ramkhamheang 592University. 593[4] 594Boyle, M.A., Kiel, K.A., 2001. A survey of 595house price hedonic studies of the impact 596of environmental externalities. Journal of 597Real Estate Literature, 9(2): 117–144. 598[5] 599Chomkrod, J. 2009. Legal measure for the 600protection and control of air pollution 601from petrochemical industrial factories: 602Case study of Map Ta Phut Industrial Es-603tate. Thesis, Thammasat University. 604