1# RECORD fb9d007b500b1eff5cd9c2ed633d123752bd42152# TITLE: EPC Labels and Building Features: Spatial Implications over Housing Prices3# YEAR: 2021.0 VENUE: nan DOI: 10.3390/su130528384# AUTHORS: Alice Barreca; E. Fregonara; D. Rolando5# PAGES: 22 CANDIDATE TABLE PAGES (1-based): [15, 17, 18]6# ABSTRACT: The influence of building or dwelling energy performance on the real estate market dynamics and pricing processes is deeply explored, due to the fact that energy efficiency improvement is one of the fundamental reasons for retrofitting the existing housing stock. Nevertheless, the joint effect produced by the building energy performance and the architectural, typological, and physical-technical attributes seems poorly studied. Thus, the aim of this work is to investigate the influence of both energy performance and diverse features on property prices, by performing spatial analyses on a sample of housing properties listed on Turin’s real estate market and on different sub-samples. In particular, Exploratory Spatial Data Analyses (ESDA) statistics, standard hedonic price models (Ordinary Least Squares—OLS) and Spatial Error Models (SEM) are firstly applied on the whole data sample, and then on three different sub-samples: two territorial clusters and a sub-sample representative of the most energy inefficient buildings constructed between 1946 and 1990. Results demonstrate that Energy Performance Certificate (EPC) labels are gaining power in influencing price variations, contrary to the empirical evidence that emerged in some previous studies. Furthermore, the presence of the spatial effects reveals that the impact of energy attributes changes in different sub-markets and thus has to be spatially analysed.789===== PAGE 1 (score 1, numeric-paren 0, table-words 0, var-words 1) =====101120 October 202612POLITECNICO DI TORINO13Repository ISTITUZIONALE14EPC Labels and Building Features: Spatial Implications over Housing Prices / Barreca, A., Fregonara, E., Rolando, D.. -15In: SUSTAINABILITY. - ISSN 2071-1050. - 13:5(2021), p. 2838. [10.3390/su13052838]16Original17EPC Labels and Building Features: Spatial Implications over Housing Prices18Publisher:19Published20DOI:10.3390/su1305283821Terms of use:22Publisher copyright23(Article begins on next page)24This article is made available under terms and conditions as specified in the corresponding bibliographic description in25the repository26Availability:27This version is available at: 11583/2873980 since: 2021-03-28T13:08:40Z28MDPI293031===== PAGE 15 (score 46, numeric-paren 38, table-words 4, var-words 0) =====3233Sustainability 2021, 13, 28383414 of 2135showed the presence of spatial dependence between variables (Moran’s I = 0.346), thus the36OLS model could be biased.37Furthermore, results showed that, above the spatial regression models, the more38suitable spatial model tested was the Spatial Error Model (SEM), since AIC and Log39likelihood values are respectively lower and higher than in the Spatial Lag Model(SLM).40The Breusch-Pagan test on the spatial effects, calculated for testing the homoscedasticity41hypothesis, showed that the null hypothesis was confirmed (Table 3).42Table 3. Ordinary Least Squared (OLS) and Spatial Error Models (SEM), complete sample (Source: Authors’ elaboration).43New Housing Listing Prices per Square44Meter (LogLP)45New Housing Listing Prices per Square46Meter (LogLP)47Variables48Ordinary Least Squares (OLS)49Spatial Error Model (SEM)50Coefficients51Probability52Coefficients53Probability54NRM55−0.009560.33157−0.021580.00159***60NBT610.070620.00063***640.034650.00166***67LTR = 068Omitted69LTR = 1700.081710.00072***730.085740.00075***76NVW770.017780.170790.031800.00181***82CSS = 083Omitted84CSS = 1850.128860.00087***880.047890.00190***91PNT = 092Omitted93PNT = 1940.092950.01496*970.090980.00099***100CBX = 0101Omitted102CBX = 11030.0191040.2931050.0571060.000107***108LFT = 0109Omitted110LFT = 11110.0691120.000113***1140.1111150.000116***117CTP = 11180.0161190.392120−0.0341210.013122*123CTP = 2124−0.1001250.000126***127−0.0701280.000129***130CTP = 3131Omitted132MTL = 1133−0.0531340.031135*136−0.0571370.001138***139MTL = 2140Omitted141MTL = 31420.0991430.000144***1450.0961460.000147***148MTL = 41490.1991500.000151***1520.1811530.000154***155BLC = 1156Omitted157BLC = 21580.1941590.000160***1610.1001620.001163***164BLC = 31650.3881660.000167***1680.1591690.000170***171BLC = 41720.6061730.000174***1750.2051760.000177***178BLC = 51790.8311800.000181***1820.2511830.000184***185BLT = 0186Omitted187BLT = 11880.0881890.005190**1910.0961920.000193***194ARC = 0195Omitted196ARC = 11970.0541980.017199*2000.0452010.004202**203EPC = 32040.0252050.5852060.0622070.050208*209EPC = 2210Omitted211EPC = 1212−0.0332130.018214*215−0.0272160.005217**218LAMBDA219-220-2210.7972220.000223***224Constant2256.9002260.0002277.1842280.000229Number of observations2302092.0002312092.000232Log likelihood233−478.61623422.744235AIC2361001.230237−1.487238R square2390.5372400.754241Adj R square2420.532243-244Breush-Pagan test24598.4712460.00024784.6932480.000249Likelihood ratio test2501002.7192510.000252DIAGNOSTICS FOR SPATIAL DEPENDENCE FOR WEIGHT MATRIX: LP_2015–2018 2092_ (row-standardized weights)253TEST254MI/DF255VALUE256PROB257Moran’s I (error)2580.34625927.1182600.000261Lagrange Multiplier (lag)2621.0002631279.5162640.000265Robust LM (lag)2661.000267560.6042680.000269Lagrange Multiplier (error)2701.000271722.4962720.000273Robust LM (error)2741.0002753.5842760.058277Lagrange Multiplier (SARMA)2782.0002791283.1002800.000281Signif codes: p ≤0.001 ‘***’; p ≤0.01 ‘**’; p ≤0.05 ‘*’; p ≤0.1 ‘.’; p ≤1 ‘ ’.282ResultsshowedthattheSEMmodelcanexplain75%ofthepricevariation(R squared = 0.754)283and regression residuals were not clustered. This result is rather good considering that284the aim of this study was not to predict property values, but just to study the influence285286287===== PAGE 16 (score 25, numeric-paren 0, table-words 4, var-words 17) =====288289Sustainability 2021, 13, 283829015 of 21291of a set of characteristics on property prices. The significant variables with the higher292marginal prices were the presence of elevator (LFT (1) = 0.111), the renovated condition of293the unit (MTL (4) = 0.181), the category of the building, that means prestigious buildings294(BLC (5) = 0.250) and the coefficient on the spatially correlated errors (LAMBDA = 0.797),295which were able to manage the spatial dependence in the model.296On the other hand, in relation with the characteristics of the housing units, a clear,297even if small, influence of the EPC label and the presence of an air conditioning system298on housing prices emerged. In particular, low EPC labels from G to E (EPC (1) = −0.027)299significantly and negatively affected housing prices, while high EPC labels from B to A4300(EPC (3) = 0.062) had a positive influence on them, and the presence of air conditioning301had a positive small marginal coefficient (ARC = 0.045). These results are in line with the302recent literature related to the perception of the EPC label in the real estate market, which303is an ongoing research field [74]. For example, Chen and Marmolejo-Duarte highlighted304that in Barcelona, an energy performance improvement from label G to label A brought in305a growth of 8.6% of housing prices in 2014, and an increase of 10.6% from label G to label B306in 2016 [75]. Moreover, Dell’Anna et al. estimated an increase given by the EPC of 6.33%307for each rating level from G to A in Turin for the 2014–2018 time period [11].308By comparing the results of the current analysis with two previous researches based309on the same study area (Turin) with the same source of data (TREMO) [12,13] and with310other recent literature [54,75,76], it is possible to confirm that the attention on the EPC311label rose slightly in recent years and is starting to be monetized by the real estate market,312not only for the lower labels (E, F, G) but also for the higher ones (A-A4, B).313This means that the international energy policies and regulations transposed into314the national laws and legal constrains (EPC is mandatory for all transaction contracts315since 2012, and minimum energy requirements are mandatory for all types of building316interventions in Turin since 2007) applied on the residential building stock in the city of317Turin are finally starting to be recognized by the real estate market.3185.2.1. Spatial Clustering: “Northern Zone” and “Central Zone” Sub-Samples319A further step was carried out by applying the SEM on both the CZ and the NZ sub-320samples, which were previously defined. The CZ sub-sample included 886 observations,321while the NZ sub-sample included 660 observations. Both spatial regressions assumed the322logarithm of unitary listing price (LogLP) as a dependent variable, while, on the basis of323the significance of the independent variables (Stepwise selection), different intrinsic and324extrinsic characteristics were included in the two final models (Table 4).325By comparing the results of the two models, the following variables were significant326both in the NZ and in the CZ sub-sample: the presence of Terrace and Lift, the Maintenance327Level, the Building Type and the 1946–1975 Construction Time Period (CTP (2)). Further-328more, the presence of Custodian Service and Air Conditioning variables were significant329only in the NZ sub-sample, while the Allocation Level, the Number of Bedrooms and330Bathrooms, the Number of Views, and the classification of the unit as a Penthouse were331significant in the CZ sub-sample. Although the Building Category represents one of the332most significant variables, there are also some differences in its significance in the two333sub-samples: in the NZ sub-sample almost all levels were significant, while in the CZ only334higher levels (Noble (4) and Prestigious (5)) were significant. The analysis of the EPC label335highlighted that the high EPC labels (3) were significant only in the CZ sub-sample, while336low EPC labels (1) had a negative influence on the price formation process only in the NZ337sub-sample.338339340===== PAGE 17 (score 51, numeric-paren 34, table-words 3, var-words 11) =====341342Sustainability 2021, 13, 283834316 of 21344Table 4. Spatial Error Model, Central Zone (CZ) and Northern Zone (NZ) sub-samples (Source: Authors’ Elaboration).345Listing Prices per Square Meter (LogLP)CENTRAL ZONE346Listing Prices per Square Meter (LogLP)NORTHERN ZONE347Variables348Spatial Error Model (SEM)349Variables350Spatial Error Model (SEM)351Coefficients352Probability353Coefficients354Probability355NRM356−0.0333570.000358***359CSS = 13600.0823610.025362*363NBT3640.0433650.001366***367CBX = 13680.0633690.025370*371LTR = 13720.0933730.000374***375LTR = 13760.1053770.001378***379NVW3800.0293810.019382**383LFT = 13840.1443850.000386***387ALL = 23880.063890.001390***391MTL = 1392−0.0933930.002394***395ALL = 33960.0543970.001398***399MTL = 34000.0894010.001402***403PNT = 14040.0964050.003406**407MTL = 44080.214090.000410***411LFT = 14120.0914130.000414***415BLT = 14160.2174170.000418***419MTL = 34200.1014210.000422***423BLC = 44240.2664250.000426***427MTL = 44280.1544290.000430***431BLC = 54320.2814330.001434***435BLT = 14360.0724370.036438*439BLC = 24400.1254410.009442**443BLC = 44440.064450.000446***447BLC = 34480.184490.000450***451BLC = 54520.1384530.000454***455CTP = 2456−0.0694570.001458**459CTP = 2460−0.0454610.015462*463EPC = 34640.0394650.544466EPC = 34670.0794680.054469.470EPC = 1471−0.0374720.049473*474EPC = 24750.0314760.034477*478ARC = 14790.0834800.014481*482ARC = 14830.0244840.245485LAMBDA4860.6274870.000488***489LAMBDA4900.654910.000492***493Constant4946.9194950.000496***497Constant4987.5814990.000500***501N of observations502886503-504N of observations505660506-507Log likelihood508119.010509-510Log likelihood511−20.489512-513AIC514−202.021515-516AIC51774.977518-519R square5200.583521-522R square5230.603524-525Breush-Pagan test52636.1435270.004528Breush-Pagan test52936.1185300.003531Likelihood ratio test532209.3385330.000534Likelihood ratio test535167.2245360.000537Signif. codes: p ≤0.001 ‘***’; p ≤0.01 ‘**’; p ≤0.05 ‘*’; p ≤0.1 ‘.’; p ≤1 ‘ ’.538In conclusion, the results confirmed that different spatial clusters behave as different539sub-markets: they both include different listing prices and building categories and are able540to differently monetize the intrinsic characteristics of the buildings.541Firstly, there is a clear connection between the impact of EPC label on housing prices542and the quality of the buildings and apartments. Namely, dwellings with a high energy-543label (EPC = A), which slightly impacts housing prices, are more expensive and boast544the best architectonic attributes. Moreover, those dwellings with high EPC labels are545located in buildings that are older than 60 years and are probably totally refurbished, both546being mainly historical buildings listed by Superintendence of Archaeology, Fine Arts547and Landscape, and being regularly under ordinary maintenance by the owners, often548representative of the richest population.549Secondly, some attributes totally differed in the two sub-samples. The characteristic550related to the air conditioning system is emblematic: it is not monetized by the city centre551sub-market as it is a common dwelling’s feature, in order to guarantee a high internal552comfort to almost all residential units (mainly classy or prestigious). On the opposite hand,553air conditioning is a stronger discriminating factor for dwelling buying decisions in the554northern suburbs, where the housing quality is lower and for which the “housing added555services” seems to compensate by acquiring the greatest importance.556The other dimensional and physical features seem related to the micro-context of the557dwelling. In fact, in the city centre the building oldness and density allow the presence of558non-ordinary apartment plans, with no balconies or with a solely principal view. Further-559more, the building density means that classy buildings are widely closed, and medium or560low floor levels have low lighting and face the opposite buildings. Thus, the presence of a561562563===== PAGE 18 (score 28, numeric-paren 12, table-words 5, var-words 6) =====564565Sustainability 2021, 13, 283856617 of 21567terrace (in the penthouse) or more than one view are appreciated features, which emerged568in previous studies [77].569Even these last considerations which emerged from the study seem to confirm the570results reported in the most recent literature (previously mentioned in the Section 2),571which reveal an uneven importance of building features—green features included—in572view of the building energy performance and, jointly, to the reference sub-market. Then,573the results can allow us to argue that, on the one side, some attributes are able to capture574a part of the EPC and/or the energy performance features, in terms of marginal prices,575for reasons that should be furtherly explored (for example, by analysing the awareness576degree of the EPC potentialities among buyers); on the other side, to argue the presence of577“latent” variables able to act as a proxy of the building/dwelling energy rating, due to the578knowledge degree of the EPC potentialities which could be surely reinforced.5795.2.2. EPC Clustering: Low Energy-Efficient Dwellings Sub-Sample580A further step of the analysis was conducted by considering the LEE sub-sample,581consisting of 234 observations related to housing units with low EPC labels (E, F and G),582built in the 1946–1990 time period. The identification of this not-spatial sub-sample, based583on an attribute selection, moved from the results of the previous spatial regressions, which584highlighted that the low EPC labels (EPC = 1) and the buildings built between 1946 and5851990 (CTP = 2), to two variables that significantly and negatively influence the listing586price formation.587A spatial regression (SEM) was performed also on this sub-sample, assuming LogLP as588the dependent variable and different intrinsic features as independent variables. The results589(final model with only significant variables) are illustrated in Table 5.590Table 5. Spatial Error Model (SEM) Low Energy Efficient (LEE) Sub-Sample. (Source: Authors’591elaboration).592New Housing Listing Prices per Square Meter (LogLP)593Variables594Spatial Error Model (SEM)595Coefficients596Probability597LTR = 15980.1495990.001600***601CSS = 16020.1466030.003604**605CBX = 16060.1146070.004608**609LFT = 16100.0906110.016612*613MTL = 36140.1166150.000616***617MTL = 46180.1896190.000620***621BLC = 36220.1286230.000624***625BLC = 46260.2146270.000628***629BLC = 56300.3876310.001632***633ARC = 16340.1436350.005636**637LAMBDA6380.7396390.000640***641Constant6427.0826430.000644***645Number of observations646234647Log likelihood6480.356649AIC65021.287651R square6520.718653Breush-Pagan test6549.5996550.476656Likelihood ratio test65781.3146580.000659Signif. codes: p ≤0.001 ‘***’; p ≤0.01 ‘**’; p ≤0.05 ‘*’; p ≤0.1 ‘.’; p ≤1 ‘ ’660Above the numerous variables and their related levels, the final SEM can explain661the 72% (R squared = 0.718) of the price formation process; results showed the spatial662LAMBDA indicator (Lambda = 0.739), the prestigious (BLC (5) = 0.387) to medium level663(BLC (3) = 0.128) of building category the high level of maintenance (MTL (4) = 0.189),664and the presence of terrace, lift, box and custodian service as the most influential variables.665666667===== PAGE 19 (score 12, numeric-paren 0, table-words 0, var-words 12) =====668669Sustainability 2021, 13, 283867018 of 21671In conclusion, in this case the spatially correlated errors indicator (expressed by672the LAMDA variable) and the “building quality” (expressed by the prestigious building673category) are able to explain the greatest part of the price variability. It is interesting to674notice that in this sub-sample of energy inefficient housing, refurbished dwellings are also675included with a good level of maintenance and prestigious building category, confirming676that the energy efficiency and the techno-physical features related to the energy behavior of677a building, in the case of existing and historical buildings, are often unrelated to the building678architectural quality. Moreover, also for the sub-sample of existing assets, the “housing679added services” such as the presence of the appurtenance car box, the presence of lift and680custodian service are attributes appreciated by the market.681Finally, it is worth highlighting the presence of two variables probably more related682with the present climate changes: the presence of terrace/balcony and the presence of air683conditioning. Since the average temperatures increased significantly in Turin from +8.07 ◦C684(in 1963) to +10.69 ◦C (in 2018) [78], devices (passive and active) related to home cooling685are increasingly appreciated by the market.6866. Conclusions687This paper aimed to explore the pricing processes of the real estate residential proper-688ties, considering the influence of the EPC labels and relevant building/dwelling attributes689on prices in their spatial context. The study was developed by analysing a sample of690residential listing prices in Turin. Exploratory Spatial Data Analyses (ESDA) and two691regression models (the Ordinary Least Squares model and the Spatial Error Model) were692firstly applied on the whole data sample.693The results highlighted that low EPC labels (E, F and G) significantly and negatively694affect housing prices, while high EPC labels (B, A1, A2, A3 and A4) have a lower but695positive influence on them. Moreover, some intrinsic building/dwelling features emerged696as characteristics particularly able to influence the property price formation: the building697category and the housing unit maintenance level.698Secondly, the regression models were applied on three different sub-samples: two of699them were generated by a spatial approach—based on LISA cluster maps of LP—while700the third was identified by applying an attribute (not spatial) approach based on intrinsic701features, representative of the most energy inefficient housing units. The results showed702that high EPC labels are significant only in the “high-high” spatial cluster characterized703by a positive autocorrelation of high LP values, while low EPC labels have a negative704influence on the price formation process only in the “low-low” spatial cluster, character-705ized by a positive autocorrelation of low LP values. By comparing the influence of the706building/dwelling attributes on property prices in these two spatial sub-samples, some707variables resulted as always significant, but several differences emerged, which means that708different spatial clusters behave as different sub-markets.709The results achieved in the third sub-sample, which included low energy efficient710dwellings, on the one hand confirmed the great influence of prestigious “building category”711and high levels of maintenance, and on the other highlighted the relevance of other712variables, such as the presence of a terrace, lift, box, custodian service and air-conditioning713system. Therefore, the results allow for arguing that the EPC labels are acquiring an714explicative power on property prices, in line with some international studies. Specifically,715by comparing the results with two previous studies conducted in the city of Turin in two716different periods [12,13], a growing effect of EPCs on prices can be seen.717Furthermore, some building features related to the energy performance emerged718which deserve to be deeply explored in future, being connected to two relevant aspects: on719one side the capacity of certain features, including the “green” ones, to capture a part of720the EPC’s explicative power; on the other side, the eventual presence of “latent” variables721acting as a proxy of the building/dwelling energy rating.722Operatively, the methodological approach illustrated in this work can be considered723a support for building/dwelling retrofitting policies, considering the potentialities of724