1# RECORD 8bd308f78fb4466f73af2035a6682d0f84ccee3e2# TITLE: Residential Property Behavior Forecasting in the Metropolitan City of Milan: Socio-Economic Characteristics as Drivers of Residential Market Value Trends3# YEAR: 2021.0 VENUE: Sustainability DOI: 10.3390/su130736124# AUTHORS: Marzia Morena; G. Cia; L. Baiardi; Juan Sebastián Rodríguez Rojas5# PAGES: 25 CANDIDATE TABLE PAGES (1-based): [15, 16, 17]6# ABSTRACT: The phenomenon of urbanization of cities has been the subject of numerous studies and evaluation protocols proposing to analyze the degree of economic and social sustainability of development projects. Through careful research and synthesis of the theoretical framework regarding residential properties’ performance measurement and forecasting, this paper goes deeper into the proposition of property development as an asset class that represents the biggest share of the Italian property market and yet is avoided by the big portfolios. The analysis model was applied to the city of Milan and its Metropolitan Area. The method is based on the development of correlation indices to evaluate different behaviors, through time and a Geographic Information System (GIS) based on the Hedonic Price Method (HPM). Results from a hedonic model estimated for several recent years suggest that, depending on the particular view, the relation between the rent/price performance and the different external and intrinsic variables can represent a useful parameter for evaluating the feasibility of different real estate investments.789===== PAGE 1 (score 6, numeric-paren 0, table-words 0, var-words 6) =====1011sustainability12Article13Residential Property Behavior Forecasting in the Metropolitan14City of Milan: Socio-Economic Characteristics as Drivers of15Residential Market Value Trends16Marzia Morena *17, Genny Cia18, Liala Baiardi19and Juan Sebastián Rodríguez Rojas202122Citation: Morena, M.; Cia, G.;23Baiardi, L.; Rodríguez Rojas, J.S.24Residential Property Behavior25Forecasting in the Metropolitan City26of Milan: Socio-Economic27Characteristics as Drivers of28Residential Market Value Trends.29Sustainability 2021, 13, 3612.30https://doi.org/10.3390/su1307361231Academic Editor: Jorge Chica-Olmo32Received: 1 February 202133Accepted: 17 March 202134Published: 24 March 202135Publisher’s Note: MDPI stays neutral36with regard to jurisdictional claims in37published maps and institutional affil-38iations.39Copyright: © 2021 by the authors.40Licensee MDPI, Basel, Switzerland.41This article is an open access article42distributed43under44the45terms46and47conditions of the Creative Commons48Attribution (CC BY) license (https://49creativecommons.org/licenses/by/504.0/).51Department of Architecture, Built Environment and Construction Engineering, Politecnico di Milano,52Via Bonardi 9, 20133 Milan, Italy; genny.cia@polimi.it (G.C.); liala.baiardi@polimi.it (L.B.);53juansebastian.rodriguez@mail.polimi.it (J.S.R.R.)54* Correspondence: marzia.morena@polimi.it; Tel.: +39-02-2399-518955Abstract: The phenomenon of urbanization of cities has been the subject of numerous studies56and evaluation protocols proposing to analyze the degree of economic and social sustainability of57development projects. Through careful research and synthesis of the theoretical framework regarding58residential properties’ performance measurement and forecasting, this paper goes deeper into the59proposition of property development as an asset class that represents the biggest share of the Italian60property market and yet is avoided by the big portfolios. The analysis model was applied to the city61of Milan and its Metropolitan Area. The method is based on the development of correlation indices62to evaluate different behaviors, through time and a Geographic Information System (GIS) based on63the Hedonic Price Method (HPM). Results from a hedonic model estimated for several recent years64suggest that, depending on the particular view, the relation between the rent/price performance65and the different external and intrinsic variables can represent a useful parameter for evaluating the66feasibility of different real estate investments.67Keywords: real estate market; residential property; hedonic price method (HPM); urban sustainabil-68ity; urban transformation691. Introduction70The nature of the property market causes the construction of predictions to be de-71pendent on many factors that are often ignored or undervalued, as well as on the already72existing factors that determine less perceptible changes. According to various international73research works carried out at an urban level, the true nature of the change in real estate74values lies in trends that may be due to the simple perception of the population [1,2], a75complex urban process [3,4], or, on occasion, situations where the price system may not76immediately reflect the changes suffered by the area in recent times [5,6].77The possibility that residential Real Estate might not be efficiently upraised was78given a strong foundation by Case and Shiller (1989) [7]. The authors state that there is a79correlation and predictability in the average cost of housing prices but establishes the idea80of a possible correlation between future prices with immediate rents as a natural behavior81of residential real estate. When considering residential properties, predictability states and82reinforces the idea of a sequential correlation [8].83The susceptibility of house prices, and the model with which such prices are calculated,84are presented by Poterba (1984) [9] stating that a perfectly informed model of house prices85may have positive shocks followed by declining prices and vice versa; this is related to the86lags in the supply of new assets. Even if these scenarios are known and incorporated into87the model, the nature of the market creates a serial correlation and then, even if these lags88are well known and fully incorporated into expectations, “they still create serial correlation89and (somewhat) predictable house price movements in reaction to shocks” [10].90Sustainability 2021, 13, 3612. https://doi.org/10.3390/su1307361291https://www.mdpi.com/journal/sustainability929394===== PAGE 15 (score 25, numeric-paren 0, table-words 2, var-words 12) =====9596Sustainability 2021, 13, 36129715 of 2598 99Figure 5. Transactions used for the construction of the model.100The model was built by a natural log transformation of the housing price (Y) and a101total of 25 variables (X). The first run with all variables can be seen below. Since a non-linear102nature of the variables is expected, a logarithmic (log-log) regression was tested (Table 3);103log-log transformation implies that coefficients are a percentage change of the variable Y to104a percentage change of the variable X.105Table 3. First run regression with all the variables.106Variable107Coeff108Coeff. EE109T Value110P Value111FIV112Constante11310.8361140.96511511.231160.000117Bathrooms1180.1201190.1141201.051210.2991223.59123Rooms124−0.23851250.0956126−2.501270.0161282.84129Old Coefficient1300.23211310.08371322.771330.0081342.24135State1360.28111370.09481382.961390.0051401.39141Status of the neighbourhood1420.12951430.06641441.951450.0571462.07147Distance to the center148−0.3691490.108150−3.411510.0011527.19153Degradation Proximity Coefficient154−0.07341550.0204156−3.591570.0011582.36159Big Park Proximity Index1600.10891610.04621622.351630.0231642.05165UrbanVegetation Proximity Index166−0.01751670.0121168−1.441690.1571701.36171Special green areas Index1720.05871730.03011741.951750.0571762.64177Equiped Green Areas Proximity Index178−0.03551790.0272180−1.301810.1991821.96183Contaminated areas Index184−0.00561850.0193186−0.291870.7721881.96189Potentially Contaminated Areas Index1900.01411910.02181920.651930.5201942.58195Metro Proximity Index1960.05871970.02231982.631990.0112002.97201Bus and Tram Proximity202−0.02032030.0296204−0.692050.4962061.74207Train Proximity Index2080.03982090.02042101.962110.0562121.77213Iconic places Proximity Index2140.05792150.03472161.672170.10221810.93219Kindergarden Proximity Index220−0.04032210.0685222−0.592230.5592245.63225226227===== PAGE 16 (score 68, numeric-paren 0, table-words 26, var-words 5) =====228229Sustainability 2021, 13, 361223016 of 25231Table 3. Cont.232Variable233Coeff234Coeff. EE235T Value236P Value237FIV238Primary schools Proximity Index239−0.17862400.0862241−2.072420.0442437.61244Secondary schools Proximity Index2450.04462460.04222471.062480.2962495.11250Universities Proximity Index2510.01962520.02482530.792540.4342554.48256C6H6257−0.01102580.0228259−0.482600.6322611.21262CO_8H2630.00032640.01492650.022660.9852671.45268NO2269−0.01482700.0320271−0.462720.6472732.44274PM10275−0.00012760.0137277−0.012780.9932791.38280where: Coeff: Percentage change of the variable Y to a percentage change of the specific variable. Coeff. EE281standard error of the coefficient. T Value: t-statistics value (ratio of the departure of the estimated value of a282parameter from its hypothesized value to its standard error). P Statistical significance value. FIV: VIF, Variance283Inflation Factor.284The first run of the regression presented an R2 adjusted value of 0.723 indicating a285‘good fit’, yet according to statistical significance rules, the variables with p value ≤0.05286prove strong presumption against the null hypothesis. Therefore, using a stepwise forward287selection, the following variables were found to be statistically significant (Table 4):288Table 4. Second run regression.289Variable290Coeff291Coeff. EE292T Value293P Value294FIV295Constante29610.8602970.75829814.332990.000300Rooms301−0.13463020.0645303−2.093040.0413051.35306Old Coefficient3070.20063080.06753092.973100.0043111.52312State3130.38113140.08253154.623160.0003171.10318Status of the neighbourhood3190.14463200.06063212.393220.0203231.80324Distance to the center325−0.40573260.0787327−5.153280.0003293.97330Degradation Proximity Coefficient331−0.06393320.0158333−4.053340.0003351.47336Big Park Proximity Index3370.09733380.03733392.613400.0113411.39342Metro Proximity Index3430.04393440.01943452.263460.0283472.35348Train Proximity Index3490.03523500.01783511.983520.0523531.41354Iconic places Proximity Index3550.06833560.02543572.693580.0093596.08360Primary schools Proximity Index361−0.10803620.0443363−2.443640.0183652.09366where: Coeff: Percentage change of the variable Y to a percentage change of the specific variable. Coeff. EE367standard error of the coefficient. T Value: t-statistics value (ratio of the departure of the estimated value of a368parameter from its hypothesized value to its standard error). P Statistical significance value. FIV: VIF, Variance369Inflation Factor.370The adjusted coefficient of determination, denoted by R2 adjusted, is 0.8274, which371suggests that 82.74% of the total variation in the Y (house price/sqm) is explained by372the relationships between Y and the selected variables. This implies that the model is373significantly ‘good fit’ (Table 5).374Table 5. Goodness of fit of the model.375S376R-Squared377R-Squared (Adjusted)378R-Squared (Pred)3790.19638238085.34%38182.74%38278.75%383The regressions results in the following:384Ln(Price €/m2) = 10,860 −0.1346 × Ln(Rooms) + 0.2006 × Ln(OldCo) + 0.3811 ×385Ln(State) + 0.1446 × Ln(NeighSt) −0.4057 × Ln(DistCBD) −0.0639×Ln(DegpCo) + 0.0973386× Ln(ParkP) + 0.0439×Ln(MetroPI) + 0.0352×Ln(TrainPI) + 0.0683×Ln(IconPI) −0.108 ×387Ln(PrsPI) + e388389390===== PAGE 17 (score 24, numeric-paren 0, table-words 5, var-words 9) =====391392Sustainability 2021, 13, 361239317 of 25394As stated by the result above, the regression analysis shows a statistically significant395premium price for the location of the asset (in particular, location in the Central Business396District) and for the parameter of dimensions, quality of the building, proximity of green397areas, metro, and primary schools.398A statistical description of the independent variables intrinsic to each transaction is399presented in the table below (Table 6).400Table 6. Statistical description of the residential intrinsic independent variables.401PRICE402[€/M2]403Rooms404Old405Coefficient406State407Status of the408Neighbour-409hood410Distance to411the Center412[m]413Calculation414Method415Given416Given4171/10 × Years418from419construction420year up to42119414221 = bad4232 = normal4243 = good4254 = New426From 1=bad427perception428up to4296= Very430Prestigious431Given432Interpretation433-434-435Lower =436Recently built437-438Mean4393.6374402.914419.814422.644433.084444.689445Standard446Error4471114480.074490.184500.044510.07452116453Median4542.9804553.004569.904573.004583.004594.373460Mode4612.2004622.004639.904643.004652.004661.872467Standard468Deviation4692.0834701.344713.354720.664731.384742.171475Coefficient476of Variation4770.704780.454790.344800.224810.464820.50483Kurtosis4841.474859.53486−0.454870.01488−0.69489−0.7490Skewness4910.674922.34493−0.03494−0.274950.354960.2497Minimum4981.2634991.005001.905011.005021.00503546504Maximum50512.43250610.0050715.705084.005096.005109.687511As can be seen from the coefficients of variation (<1.0) of each intrinsic variable, the512sample used manages to represent a general picture of all the property types that are513generally transacted in the city during Q3 2019 without biases of location and/or physical514characteristics; as can be seen furthermore, the skewness values allow to understand that515the variables considered adequately represent the whole spectrum of property quality, year516of construction, and location, represented by the distance to the city center. In the case of517number of rooms, it is clear that the number of rooms in Milan’s residential properties is518considerably asymmetrical, with an average of three rooms.519In order to understand the distribution of the spatial variables that are statistically520significant and therefore considered in the regression, a synthesis of the calculation method521is presented in Table 7.522Table 7. Calculation method of each spatial variable.523Index524Calculation Method525Degradation Proximity Coefficient526Σ of Area of degraded areas over distance to each specific527transaction in km (at least 200 sqm within 2 km).528Big Park Proximity529Σ of Area of urban parks over distance to each specific530transaction in km (at least 100 sqm within 3 km).531Metro Proximity53210 × Σ of number of metro stations to the specified transactions533over distance to each specific transaction in km (within 1.5 km).534Train Proximity53510 × Σ of number of train stations close to the transaction over536distance to each specific transaction in km (within 1 km).537Iconic Places Proximity53810 × Σ of number of monuments and historical landmarks over539distance to each specific transaction in km (within 1.5 km).540Primary Schools Proximity541Σ of number of primary schools over distance to each specific542transaction in km (within 1 km).543544545===== PAGE 18 (score 15, numeric-paren 0, table-words 4, var-words 7) =====546547Sustainability 2021, 13, 361254818 of 25549A statistical description of the independent spatial variable is presented in the table550below (Table 8):551Table 8. Statistical description of the spatial independent variables.552Degradation553Proximity554Coefficient555Big Park556Proximity557Index558Metro559Proximity560Index561Train562Proximity563Index564Iconic Places565Proximity Index566Primary Schools567Proximity Index568Mean569863.35700.757144.057210.2573193.85749.8575Standard Error57667.15770.05782.85790.858019.55810.3582Median583259.35840.658533.95869.058735.45889.9589Mode5900.05910.65920.05937.25940.05950.0596Standard Deviation597125.05980.559925.360015.0601365.86026.3603Coefficient of Variation6040.486050.836060.756071.6760810.326090.64610Kurtosis6110.76122.16130.96143.96156,46168.1617Skewness6181.16191.76201.36211.86222.66231.9624Minimum6250.06260.06270.06280.06290,06300.0631Maximum6325657.26332.463490.563574.66361681.363742.3638As can be seen from the coefficients of variation, a strong aggregation of iconic places639is evident, yet it has statistical significance, due to the fact that is not directly correlated with640the spatial distance to the city center, and due to this aggregation, several transactions are641not affected by iconic places. The remarkable city center public transportation connectivity642is well represented, and yet the sample with transactions covering different neighborhoods643is not affected by a particular aggregation of Metro stations in one single area.644The following map shows the general price situation of the used transactions. It is645important to note that the map represents the data used to build up the model, and not the646average price in the city’s areas (Figure 6). It represents a snapshot of the prices per sqm647for the 352 properties used to build the model; as can be seen, the transactions located in648the city center and its business district (Above Brera area) presented at the time span of649data collection (Q3 2019) the highest prices in the city. Some areas outside the historical650center present a prime price, such as the Tortona area and Tre Torri.651 652Figure 6. Prices of the transactions used for the model.653In order to geographically check the precision of the model, the resulting regression654was used for recalculation the prices of the transactions used to build the model, and this655