1# RECORD 942f0ff361a49b68c1333c31381f78c02a3d09372# TITLE: How different are dwellings whose energy efficiency impacts price formation?3# YEAR: 2019.0 VENUE: IOP Conference Series: Materials Science and Engineering DOI: 10.1088/1757-899x/603/4/0420154# AUTHORS: Ai Chen; Carlos Marmolejo-Duarte5# PAGES: 11 CANDIDATE TABLE PAGES (1-based): [7]6# ABSTRACT: Housing energy-efficiency has become a relevant issue since it is mandatory to exhibit an Energy Performance Certificate (EPC) when transacting real estate in the European Union. A number of studies have focused on energy-efficiency marginal prices using hedonic models from cross-sectional and longitudinal perspectives. Some of them have found that the increase of relevance of EPC ranks (i.e. marginal prices) is not the same for all the A-G classes. This study aims to explore the differences in terms of architectonic and location attributes between the apartments depicting an increase of EPC ranking marginal prices regarding those where energy efficiency seems not to play a role in price formation. In doing so, a pooled spatial error hedonic model is done using selling information for multifamily housing in Barcelona for the years of 2014 and 2016. Furthermore, it is shown which EPC ranks do exhibit an increase in terms of marginal prices in the period studied. Finally, we compare the architectonic and locative attributes for the set of homes where the energy efficiency has increased in terms of price importance to the set of homes where it has not increased. The results suggest that dwellings with high and medium EPC ranks (e.g. A for 2014; and B, C and D for 2016) are more expensive, larger and boasting of better architectonic qualities than the rest of homes where EPC ranks failed to have a role in price formation. On the contrary, the location attributes are different: while A-ranked dwellings of 2014 are located on peripheries where new housing completions are placed; B, C and D-ranked homes of 2016 are located in more centric locations. These findings have implications for future analysis regarding the energy premium and energy poverty, since specific characteristics in different submarkets may have a different impact on housing prices.789===== PAGE 1 (score 3, numeric-paren 0, table-words 0, var-words 3) =====1011IOP Conference Series:12Materials Science and13Engineering14 15PAPER • OPEN ACCESS16How different are dwellings whose energy17efficiency impacts price formation?18To cite this article: Ai Chen and Carlos Marmolejo-Duarte 2019 IOP Conf. Ser.: Mater. Sci. Eng. 6031904201520 21View the article online for updates and enhancements.22You may also like23The improved anode performance enabled24by Ni2P@C embedded in echinus-like25porous carbon for lithium-ion battery26Zhuo Chen and Haibo Li27-28Enhancement of electron–phonon29coupling due to increased magnetism and30applied hydrostatic pressure in FeSe31Xiao-Fang Ouyang, Ze-Yi Song and Yu-32Zhong Zhang33-34Improving the performance of quantum35well solar cells with photonic crystal36Weiye Liu, Jiaping Guo, Ding Ding et al.37-38This content was downloaded from IP address 132.203.214.169 on 20/10/2026 at 20:22394041===== PAGE 7 (score 38, numeric-paren 0, table-words 2, var-words 23) =====4243WMCAUS 201944IOP Conf. Series: Materials Science and Engineering 603 (2019) 04201545IOP Publishing46doi:10.1088/1757-899X/603/4/04201547648EPC rankings (ranked “A” in 2014 while “B”, “C” and “D” in 2016) have a significant impact on housing 49price formation. 50Table 2. Spatial Error Estimation Results 51Spatial Error Model - Maximum Likelihood Estimation 52R Square 530.764 54Log-likelihood 55-928 56 57Sigma Square 580.075 59AIC 601917 61 62S.E of regression 630.274 64 65 66 67 68 69 70 71 72 73Coefficients 74 75B 76Std. Error 77Z-Value 78Prob. 79Lambda 80 0.462 81 0.016 82 28.317 83 - 84(Constant) 85 10.125 86 0.056 87 182.398 88 - 89Year 2016 90 0.041 91 0.015 92 2.684 93 0.007 94Structural architectonic characteristics of apartments and buildings 95Area 96 0.015 97 0.000 98 37.546 99 - 100Area^2 101- 0.000 102 0.000 103- 19.709 104 - 105Number of bathrooms 106 0.098 107 0.008 108 11.842 109 - 110Air conditioner 111 0.074 112 0.008 113 8.905 114 - 115Central heating 116 0.041 117 0.009 118 4.686 119 - 120Retrofitted apartment 121 0.034 122 0.010 123 3.437 124 0.001 125Swimming pool 126 0.186 127 0.015 128 11.976 129 - 130Lift 131 0.089 132 0.008 133 10.452 134 - 135Inverse of building age 136 0.238 137 0.045 138 5.340 139 - 140Energy performance of apartments 141Energy class A 142 0.086 143 0.034 144 2.492 145 0.013 146Energy class C 147- 0.011 148 0.029 149- 0.375 150 0.708 151Energy class D 152 0.000 153 0.019 154 0.020 155 0.984 156Energy class E 157 0.007 158 0.013 159 0.572 160 0.567 161Energy class F 162 0.019 163 0.016 164 1.132 165 0.258 166Energy class A * Year 2016 167 0.067 168 0.044 169 1.515 170 0.130 171Energy class B * Year 2016 172 0.107 173 0.041 174 2.619 175 0.009 176Energy class C * Year 2016 177 0.106 178 0.041 179 2.597 180 0.009 181Energy class D * Year 2016 182 0.105 183 0.026 184 4.071 185 - 186Energy class E * Year 2016 187 0.008 188 0.018 189 0.462 190 0.644 191Energy class F * Year 2016 192- 0.033 193 0.024 194- 1.386 195 0.166 196Locative attributes (transport, centrality and amenities) 197Commuting time 198 0.004 199 0.001 200 2.597 201 0.009 202<200m from sea shore 203 0.131 204 0.029 205 4.532 206 - 207Highway ramp 208 0.081 209 0.022 210 3.712 211 - 212<800 m from railway station 213 0.033 214 0.011 215 3.120 216 0.002 217Centrality index 218 0.025 219 0.003 220 8.680 221 - 222Average gross area floor ratio 223 0.048 224 0.005 225 9.029 226 - 227Socio-economic attributes 228PC1 High income 229 0.104 230 0.006 231 17.356 232 - 233PC2 Med income 234 0.072 235 0.007 236 10.979 237 - 238 Note: independent variables/covariates are introduced using the stepwise method. Energy class G is the controlled group. ‘–’ indicated the 239significant is less than 0.000. Variables with grey colour are insignificant at 90% of confidence. Source: Own elaboration. 2404.2. How different are the dwellings which energy efficiency gains relevance in price formation? 241In order to explore significant differences in terms of achitectonic and locative attributes first the 6,492 242aparments are clustered in 4 groups. Group 1 (N = 73) is for the energy-labelled dwellings (ranking A) 243which impact significantly on housing prices in 2014 while the others are for Group 2 (N = 3,167) where 244no impact was found in the same year. Dwellings in the 2016 dataset are groupped in the same way, 245resulting in Group 3 (ranking B, C, D) where EPC ranking was found to have an impact on price 246formation and Group 4 where no impact was found- Each of such gropus are formed by 485 and 2,718 247apartments respectively. Next, an ANOVA test (at 90% confidence level) is implemented among these 248groups) for each of the architectonic and locative attributes. 2494.2.1. The specific characteristic performance in year periods. According to the first column of table 3, 250in the 2014 dataset all the architectonic attributes (excluded the number of bedrooms) are significantly 251different between homes clustered in Group 1 and 2. Conversely, in the remainder of accessibility and 252253254===== PAGE 8 (score 58, numeric-paren 0, table-words 1, var-words 27) =====255256WMCAUS 2019257IOP Conf. Series: Materials Science and Engineering 603 (2019) 042015258IOP Publishing259doi:10.1088/1757-899X/603/4/0420152607261socioeconomic dimensions just the centrality index, population density and employment show 262significant differences. The average unit price for an “A”-ranked dwelling with a 90 square meters’ size 263is 2,208 euros per square meter in 2014, 300 euros more per square meter than that in Group 2. Similarly, 264physical attributes (e.g. number of bathrooms, area of outdoor spaces and living room) and amenities 265(e.g. air conditioning and heating) are more present in Group 1 where the probability to find a heated 266dwelling doubles in Group 2 in relation to Group 1. Regarding the accessibility and socioeconomic 267dimensions, the “A”- ranked dwellings in 2014 are located in metropolitan peripheries where the average 268centrality index is 10.73, less than the referenced Group 2. The same is true for population and 269employment densities. All in all, energy-efficient homes in 2014 are basically located at peripheral zones 270where buildings are constructed under newer construction codes requiring efficient thermal 271performances. 272Table 3. Statistical Description for the groups in 2014 and 2016 273Variables 2742014 2752016 276Group 1 277(N=73) 278Group 2 279(N=3,167) 280ANOVA TEST 281Group 3 282(N=485) 283Group 4 284(N=2,718) 285ANOVA TEST 286F 287Sig. 288F 289Sig. 290Architectonic Attributes 291 292 293 294 295 296 297 298 299Total_price 300 197,784 162,047 30111.55 3020.001 275,192 221,404 30351.58 3040.000 305Unit_price 306 2,209 1,908 30714.87 3080.000 3,029 2,514 30967.04 3100.000 311Superficie 31289.84 31383.94 3143.20 3150.074 31690.64 31786.40 3187.37 3190.007 320No_bedrooms 3212.89 3222.92 3230.06 3240.801 3252.82 3262.89 3272.21 3280.137 329No_bathrooms 3301.52 3311.29 33214.82 3330.000 3341.50 3351.33 33638.34 3370.000 338Dum_air_conditioning 3390.67 3400.30 34147.38 3420.000 3430.60 3440.45 34537.82 3460.000 347Dum_heat 3480.92 3490.42 35075.30 3510.000 3520.81 3530.66 35441.27 3550.000 356Dum_reform 3570.30 3580.10 35930.71 3600.000 3610.28 3620.18 36330.19 3640.000 365Dum_lift 3660.86 3670.46 36848.34 3690.000 3700.79 3710.65 37236.59 3730.000 374Ages of buildings 37530.84 37645.55 37749.25 3780.000 37940.70 38046.45 38121.66 3820.000 383Construction_before 1981 3840.51 3850.84 38657.91 3870.000 3880.62 3890.77 39048.44 3910.000 392Construction_between 1982-2006 3930.25 3940.12 39511.16 3960.001 3970.21 3980.17 3994.35 4000.037 401Construction_after 2007 4020.25 4030.04 40468.77 4050.000 4060.17 4070.06 40869.12 4090.000 410Storied 4112.47 4122.14 4132.82 4140.093 4152.11 4162.19 4170.46 4180.497 419Areas_outdoor 42012.60 4219.40 4223.63 4230.057 4249.71 4259.45 4260.06 4270.800 428Areas_living 42915.18 43012.11 4316.75 4320.009 43313.19 43411.69 4356.18 4360.013 437Dum_grand_terrance 4380.18 4390.07 44013.35 4410.000 4420.12 4430.14 4440.48 4450.487 446Dum_swim_pool 4470.08 4480.04 4493.02 4500.082 4510.15 4520.10 4538.03 4540.005 455Dum_gard 4560.27 4570.09 45830.77 4590.000 4600.20 4610.15 4627.17 4630.007 464Accessibility Attributes 465 466 467 468 469 470 471 472 473Time_commuting 47424.61 47524.00 4761.32 4770.250 47824.57 47924.65 4800.19 4810.666 482Dum_sea (in 200 meters) 4830.03 4840.01 4851.49 4860.223 4870.04 4880.04 4890.26 4900.608 491Dum_highway 4920.96 4930.93 4941.18 4950.278 4960.94 4970.94 4980.10 4990.758 500Dum_trans_stations 5010.53 5020.50 5030.36 5040.547 5050.57 5060.56 5070.33 5080.568 509Index_Central 51010.73 51111.39 5125.52 5130.019 51412.47 51511.94 51616.39 5170.000 518Ratio_floor_areas 5191.84 5201.99 5210.87 5220.350 5232.43 5242.33 5251.54 5260.214 527Socioeconomic Attributes 528 529 530 531 532 533 534 535 536Proportion of university degree 5379.77 53810.88 5391.49 5400.222 54115.85 54214.21 54311.75 5440.001 545Density of population 546 16,884 22,051 5473.70 5480.054 21,623 24,730 5497.42 5500.006 551Density of employment 552 7,021 9,569 5534.89 5540.027 9,456 10,742 5556.77 5560.009 557PCA High income 558-0.23 559-0.11 5601.64 5610.201 5620.32 5630.13 56413.32 5650.000 566PCA Med income 567-0.22 568-0.24 5690.03 5700.865 5710.05 5720.01 5730.73 5740.393 575Pr_Manager 5767.28 5777.90 5781.40 5790.237 58010.26 5819.39 58210.92 5830.001 584Pr_professiones 5859.95 58611.01 5871.33 5880.249 58915.67 59014.14 59110.55 5920.001 593Pr_technics 59413.31 59513.29 5960.00 5970.974 59814.97 59914.35 6009.60 6010.002 602Pr_admin 60311.14 60410.98 6050.25 6060.614 60711.32 60811.31 6090.01 6100.935 611Pr_commer 61215.20 61315.05 6140.15 6150.702 61614.21 61714.77 6188.48 6190.004 620Pr_agricultura_fisher 6210.67 6220.75 6230.64 6240.425 6250.67 6260.69 6270.33 6280.568 629Pr_craftman 63018.06 63117.47 6320.62 6330.430 63413.69 63514.82 63611.51 6370.001 638Pr_operation 63913.73 64013.28 6410.38 6420.535 64310.22 64410.88 6454.69 6460.030 647Pr_unquality 64810.61 64910.17 6500.68 6510.408 6528.89 6539.55 6548.54 6550.004 656Notes: Variables with grey colour are insignificant at 90% of confidence. Source: Own elaboration. 657In the 2016 dataset for Groups 3 and 4, the same significant differences in the “architectonic attributes 658dimension” which are found in 2014, are identified but the number of storey, area of outdoor spaces and 659