# RECORD 2799534456858542797f23b6cf17bd1f12439819 # TITLE: Towards a Hedonic Pricing Method for the Bucharest Private Housing Market # YEAR: 2019.0 VENUE: Scientific Annals of Economics and Business DOI: 10.47743/saeb-2019-0031 # AUTHORS: Ion Radu Zilișteanu; R. Muntean; Ș. Gherghina; G. Vintilă; T. Barbu # PAGES: 25 CANDIDATE TABLE PAGES (1-based): [11, 14, 15, 16, 17, 18, 19, 20, 21] # ABSTRACT: This paper aims at exploring the drivers of dwellings’sale prices in Bucharest, Romania, over the period 2014-2018. Several housing structural attributes are covered, such as the number of rooms, useful and constructed surface, type of comfort, floor, the number of bathrooms, balconies and parking, the seniority from construction, as well as from renovation, structure type, height regime, and the duration until completion of the real estate transaction. By estimating a standard hedonic price model via OLS regression for a sample of 765 transactions, we notice that all the selected variables, except the floor level and seniority from construction, positively influence the property prices. However, in case of useful and constructed surface, nonlinear relationships with property prices were acknowledged. Robustness checks in form of quantile regressions reinforce the empirical findings. ===== PAGE 1 (score 12, numeric-paren 0, table-words 0, var-words 12) ===== Scientific Annals of Economics and Business 66 (3), 2019, 389-413 DOI: 10.2478/saeb-2019-0031 Towards a Hedonic Pricing Method for the Bucharest Private Housing Market Ion Radu Zilișteanu*, Răzvan Muntean**, Ștefan Cristian Gherghina***, Georgeta Vintilă § *, Teodora Cristina Barbu°* Abstract This paper aims at exploring the drivers of dwellings’sale prices in Bucharest, Romania, over the period 2014-2018. Several housing structural attributes are covered, such as the number of rooms, useful and constructed surface, type of comfort, floor, the number of bathrooms, balconies and parking, the seniority from construction, as well as from renovation, structure type, height regime, and the duration until completion of the real estate transaction. By estimating a standard hedonic price model via OLS regression for a sample of 765 transactions, we notice that all the selected variables, except the floor level and seniority from construction, positively influence the property prices. However, in case of useful and constructed surface, nonlinear relationships with property prices were acknowledged. Robustness checks in form of quantile regressions reinforce the empirical findings. Keywords: hedonic regression; OLS; quantiles; structural attributes. JEL classification: C21; R30. 1. INTRODUCTION The pecuniary value of nonmarket goods and services may be estimated via multifarious ways such as the contingent valuation method, the travel cost method, or the hedonic pricing method (Dahal et al., 2019). Among them, the hedonic pricing method is widely used (Selim, 2009; Z. Z. Huang et al., 2017), having as the main supposition that a homebuyer does not only * Department of Finance, Faculty of Finance and Banking, The Bucharest University of Economic Studies, Romania; e-mail: rzilisteanu@gmail.com (corresponding author). ** SC RUN IT SOLUTIONS SRL; e-mail: razvan@run.ro. *** Department of Finance, Faculty of Finance and Banking, The Bucharest University of Economic Studies, Romania; e-mail: stefan.gherghina@fin.ase.ro. § Department of Finance, Faculty of Finance and Banking, The Bucharest University of Economic Studies, Romania; e-mail: vintilageorgeta@yahoo.fr. ° Department of Money and Banking, Faculty of Finance and Banking, The Bucharest University of Economic Studies, Romania; e-mail: teodorabarbu65@yahoo.com. ===== PAGE 11 (score 20, numeric-paren 5, table-words 4, var-words 5) ===== Scientific Annals of Economics and Business, 2019, Volume 66, Issue 3, pp. 389-413 399 Variables Value Count Cumulative - Count Percent Cumulative - Percent 1 126 765 16.47059 100 Missing 0 765 0 100 Height_regime_6 0 725 725 94.77124 94.7712 1 40 765 5.22876 100 Missing 0 765 0 100 Height_regime_7 0 736 736 96.20915 96.2092 1 29 765 3.79085 100 Missing 0 765 0 100 Height_regime_8 0 742 742 96.99346 96.9935 1 23 765 3.00654 100 Missing 0 765 0 100 Height_regime_9 0 630 630 82.35294 82.3529 1 135 765 17.64706 100 Missing 0 765 0 100 Height_regime_10 0 728 728 95.1634 95.1634 1 37 765 4.8366 100 Missing 0 765 0 100 Source: Authors’ computations The output related to independent samples t test towards the sale price and structural attributes is exhibited in Table no. 4. We notice that there is a significant difference in price between apartments that are comfort 1 and those comfort 2. Therewith, significant differences in price are registered between apartments showing height regime Ground Floor + 10 Floors and the rest of dwellings. Similar statements are valid for Ground Floor + 2 Floors, Ground Floor + 5 Floors and, Ground Floor + 6 Floors. Table no. 5 exposes the results of independent samples t test regarding the duration until transaction completion and structural attributes. Thus, there is a statistically significant difference between the mean duration of apartments comfort 1 and those comfort 2. In the same vein, significant differences in duration occur between dwellings with height regime Ground Floor + 10 Floorsand the rest of apartments. Analogous remarks are true for Ground Floor + 3 Floors and Ground Floor + 5 Floors. The correlations between variables are pointed out in Table no. 6. Thereby, sale price is highly correlated with the useful (0.73) and constructed surface of the apartment (0.74). Likewise, we notice high correlations between the number of rooms and useful surface (0.80), the number of rooms and the constructed surface (0.75), as well as between the useful surface and constructed area (0.97). As such, the previously mentioned variables will be employed in distinct regression equations in order to avoid the multicolinearity issue. 4.2 The outcomes of hedonic pricing regressions The impact of each housing structural attributes on the sale price is reported in Table no. 7. Thus, most of the housing structural characteristic exert a positive influence on the housing price. The comfort exhibits the highest positive influence on price (Eq. 4), but the lowest impact is registered in case of surface, both useful (Eq. 2) and constructed (Eq. 3). ===== PAGE 12 (score 4, numeric-paren 0, table-words 2, var-words 0) ===== 400 Zilișteanu, I. R., Muntean, R., Gherghina, Ș. C., Vintilă, G., Barbu, T. C. Table no. 4 – The outcomes of independent samples t test regarding the sale price and structural attributes (2014-2018) Source: Authors’ computations Table no. 5 – The outcomes of independent samples t test regarding the duration until transaction completion and structural attributes (2014-2018) Source: Authors’ computations ===== PAGE 14 (score 70, numeric-paren 49, table-words 1, var-words 11) ===== 402 Zilișteanu, I. R., Muntean, R., Gherghina, Ș. C., Vintilă, G., Barbu, T. C. The number of rooms show a positive impact (Eq. 1) as in Selim (2009), Baranzini and Schaerer (2011). Likewise, if useful surface and constructed area increases by 1%, price will increase by 0.01%, the association being alike previous studies (Wilhelmsson, 2002; Keskin, 2008; Jim and Chen, 2009; Selim, 2009; Chen and Jim, 2010; Conroy and Milosch, 2011; Hui et al., 2012; Efthymiou and Antoniou, 2013; Z. Z. Huang et al., 2017; Wu et al., 2017; Yang et al., 2019). In addition, the number of bathrooms positively influence sale price (Eq. 6) alike D Dahal et al. (2019), Mulley et al. (2016), Mulley et al. (2018), Tian et al. (2017), Conroy et al. (2013), Conroy and Milosch (2011), Larsen and Blair (2014), Ibeas et al. (2012), alongside the number of parking (Eq. 8) similar Mulley et al. (2018) and the number of balconies (Eq. 7). The effect of structure on price is positive (Eq. 11), opposite to Selim (2009). The floor level shows a negative impact on sale price (Eq. 5), contrary to Jim and Chen (2009), Chen and Jim (2010), Baranzini and Schaerer (2011), Hui et al. (2012), Ibeas et al. (2012), Efthymiou and Antoniou (2013), Wu et al. (2017) . The number of years since construction exhibit a negative impact on price (Eq. 9), as in earlier studies (Wilhelmsson, 2002; Jim and Chen, 2009; Selim, 2009; Hui et al., 2012; Conroy et al., 2013; Czembrowski and Kronenberg, 2016; Tian et al., 2017; Wen et al., 2017b; Brecard et al., 2018; Lu, 2018; Laszkiewicz et al., 2019; Li et al., 2019; Yang et al., 2019). In case of the number of years since renovation, the impact is not statistically significant (Eq. 10), as opposed to Zhang and Leonard (2014). The highest goodness of fit is registered in case of Eq. 1 and Eq. 2, whereas the lowest explanatory power is depicted by Eq. 10 and Eq. 11. Table no. 7 – OLS regressions’ results regarding the impact of each housing structural attributes on the sale price (duration until transaction completion not included) Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) No_rooms 0.31*** (22.75) S_useful 0.01*** (36.13) S_constructed 0.01*** (35.20) Comfort 0.52*** (11.05) Floor -0.02*** (-4.26) No_bathrooms 0.49*** (17.23) No_balconies 0.29*** (8.71) No_parking 0.30*** (4.74) Seniority_construction -0.01*** (-7.60) Seniority_renovation -0.00 (-1.23) Structure 0.08† (1.65) _cons 10.29*** 10.12*** 10.17*** 10.50*** 11.11*** 10.40*** 10.70*** 10.79*** 11.21*** 11.00*** 10.92*** (304.83) (386.15) (399.40) (235.16) (393.12) (277.26) (259.13) (169.27) (357.95) (471.73) (230.38) F statistic 517.48*** 1305.34*** 1239.18*** 122.17*** 18.16*** 296.74*** 75.84*** 22.44*** 57.76*** 1.50 2.72† R-sq 0.40 0.63 0.62 0.14 0.02 0.28 0.10 0.07 0.08 0.00 0.00 # 765.00 760.00 762.00 734.00 765.00 765.00 706.00 317.00 685.00 451.00 669.00 Source: Authors’ computations.† p < 0.10. * p < 0.05. ** p < 0.01. *** p < 0.001 ===== PAGE 15 (score 102, numeric-paren 68, table-words 7, var-words 10) ===== Scientific Annals of Economics and Business, 2019, Volume 66, Issue 3, pp. 389-413 403 The impact of every housing structural feature on the sale price, when controlling for the duration until transaction completion, is pointed out in Table no. 8. The results are matching with those displayed in Table no. 7. Table no. 8 – OLS regressions’ results regarding the impact of each housing structural attributes on the sale price(duration until transaction completion included) Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) No_rooms 0.30*** (22.31) S_useful 0.01*** (35.48) S_constructed 0.01*** (34.56) Comfort 0.51*** (10.72) Floor -0.02*** (-4.23) No_bathrooms 0.48*** (17.01) No_balconies 0.28*** (8.47) No_parking 0.28*** (4.39) Seniority_construction -0.01*** (-6.98) Seniority_renovation -0.00 (-1.32) Structure 0.10* (1.97) Duration 0.00* -0.00 -0.00 0.00 0.00*** 0.00*** 0.00** 0.00† 0.00** 0.00** 0.00*** (2.45) (-0.26) (-0.79) (1.56) (4.19) (3.55) (2.73) (1.67) (3.05) (3.05) (4.20) _cons 10.27*** 10.12*** 10.17*** 10.50*** 11.06*** 10.37*** 10.68*** 10.77*** 11.16*** 10.95*** 10.86*** (301.99) (384.10) (397.60) (234.22) (362.92) (273.54) (255.30) (167.85) (313.43) (401.74) (220.55) F statistic 263.46*** 651.90*** 619.59*** 62.43*** 18.05*** 156.90*** 41.99*** 12.68*** 33.89*** 5.40** 10.24*** R-sq 0.41 0.63 0.62 0.15 0.05 0.29 0.11 0.07 0.09 0.02 0.03 # 765.00 760.00 762.00 734.00 765.00 765.00 706.00 317.00 685.00 451.00 669.00 Source: Authors’ computations.† p < 0.10. * p < 0.05. ** p < 0.01. *** p < 0.001 Table no. 9 and Table no. 10 exposes the effect of each height regime on the sale price. We notice mixed relationships, namely negative in case of the Ground Floor + 10 Floors (Eq. 1), but positive in case of Ground Floor + 11 Floors (Eq. 2), Ground Floor + 2 Floors (Eq. 3), Ground Floor + 5 Floors (Eq. 6), Ground Floor + 6 Floors (Eq. 7), Ground Floor + 8 Floors (Eq. 9). As well, the regression coefficients do not differ when controlling for duration until transaction completion). Table no. 11 and Table no. 12 exhibits the impact of the entire range of housing structural attributes on the sale price. Similar to previous estimations, the positive impact related to the number of rooms (Eq. 1 and Eq. 2), comfort (Eq. 1 – Eq. 10), number of bathrooms (Eq. 1, Eq. 7 and, Eq. 8) and parking (Eq. 1, Eq. 2, Eq. 4 – Eq. 6, Eq. 8 and, Eq. 10), alongside structure (Eq. 3, Eq. 7, Eq. 9 and, Eq. 10), maintains its statistical significance. Therewith, the negative influence of the floor level (Eq. 1, Eq. 4, Eq. 6, Eq. 8 and, Eq. 10) and seniority from construction (Eq. 1, Eq. 2, Eq. 6 and, Eq. 10) is strengthened. However, the effect of the number of balconies (Eq. 7 and Eq. 9) and seniority from renovation (Eq. 4) turn out to be negative, but the statistical significance is week. ===== PAGE 16 (score 126, numeric-paren 106, table-words 2, var-words 0) ===== 404 Zilișteanu, I. R., Muntean, R., Gherghina, Ș. C., Vintilă, G., Barbu, T. C. Table no. 9 – OLS regressions’ outcomes towards the influence of each height regime on the sale price (duration until transaction completion not covered) Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) Height_regime_1 -0.23*** (-7.75) Height_regime_2 0.18* (2.06) Height_regime_3 0.30** (3.09) Height_regime_4 -0.01 (-0.23) Height_regime_5 0.04 (1.04) Height_regime_6 0.23*** (3.37) Height_regime_7 0.17* (2.19) Height_regime_8 0.06 (0.71) Height_regime_9 0.13*** (3.38) Height_regime_10 -0.05 (-0.66) _cons 11.09*** 11.00*** 11.00*** 11.01*** 11.00*** 10.99*** 11.00*** 11.00*** 10.98*** 11.01*** (617.04) (725.05) (729.38) (704.72) (671.37) (719.60) (722.30) (723.52) (670.13) (716.89) F statistic 60.04*** 4.24* 9.54** 0.05 1.08 11.35*** 4.79* 0.50 11.45*** 0.43 R-sq 0.07 0.01 0.01 0.00 0.00 0.01 0.01 0.00 0.01 0.00 # 765.00 765.00 765.00 765.00 765.00 765.00 765.00 765.00 765.00 765.00 Source: Authors’ computations.† p < 0.10. * p < 0.05. ** p < 0.01. *** p < 0.001 Table no. 10 – OLS regressions’ outcomes towards the influence of each height regime on the sale price (duration until transaction completion covered) Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) Height_regime_1 -0.22*** (-7.35) Height_regime_2 0.18* (2.04) Height_regime_3 0.27** (2.87) Height_regime_4 -0.05 (-0.99) Height_regime_5 0.04 (1.09) Height_regime_6 0.19** (2.78) Height_regime_7 0.17* (2.21) Height_regime_8 0.07 (0.86) Height_regime_9 0.14*** (3.63) Height_regime_10 -0.03 (-0.48) Duration 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** (3.48) (4.21) (4.06) (4.32) (4.23) (3.76) (4.23) (4.24) (4.42) (4.19) _cons 11.05*** 10.95*** 10.95*** 10.96*** 10.95*** 10.95*** 10.95*** 10.95*** 10.93*** 10.96*** (503.69) (573.98) (578.31) (573.84) (544.66) (578.05) (572.31) (570.65) (541.08) (566.22) F statistic 36.53*** 11.02*** 13.09*** 9.38*** 9.49*** 12.84*** 11.39*** 9.25*** 15.63*** 9.00*** R-sq 0.09 0.03 0.03 0.02 0.02 0.03 0.03 0.02 0.04 0.02 # 765.00 765.00 765.00 765.00 765.00 765.00 765.00 765.00 765.00 765.00 Source: Authors’ computations.† p < 0.10. * p < 0.05. ** p < 0.01. *** p < 0.001 ===== PAGE 17 (score 200, numeric-paren 158, table-words 1, var-words 7) ===== Scientific Annals of Economics and Business, 2019, Volume 66, Issue 3, pp. 389-413 405 Table no. 11 – The output of OLS regressions regarding the impact of the entire range of housing structural attributes on the sale price (duration until transaction completion not included) Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) No_rooms 0.23*** 0.24*** (9.19) (9.95) Comfort 0.34*** 0.31*** 0.26*** 0.14* 0.20** 0.12* 0.26*** 0.15* 0.21** 0.12* (5.79) (5.21) (4.33) (2.54) (3.32) (2.07) (4.28) (2.51) (3.26) (2.06) Floor -0.01* -0.01 -0.01 -0.01** -0.01 -0.01* -0.01 -0.01* -0.01 -0.01* (-2.03) (-1.49) (-1.45) (-2.66) (-1.37) (-2.42) (-1.20) (-2.09) (-1.35) (-2.19) No_bathrooms 0.09† 0.03 0.07 0.06 0.04 0.03 0.09† 0.08† 0.06 0.05 (1.92) (0.70) (1.52) (1.41) (0.90) (0.80) (1.80) (1.85) (1.14) (1.19) No_balconies -0.03 -0.04 -0.05 -0.02 -0.06 -0.03 -0.08† -0.05 -0.09† -0.05 (-0.71) (-0.89) (-1.12) (-0.60) (-1.41) (-0.82) (-1.81) (-1.21) (-1.97) (-1.30) No_parking 0.21** 0.23** 0.09 0.16* 0.13† 0.19** 0.06 0.16* 0.09 0.16* (3.05) (3.15) (1.35) (2.53) (1.71) (2.83) (0.82) (2.43) (1.12) (2.23) Seniority_construction -0.00*** -0.01*** -0.00 -0.00 -0.00 -0.00† -0.00 -0.00 -0.00 -0.00* (-3.76) (-4.57) (-0.43) (-1.20) (-0.89) (-1.83) (-1.03) (-1.55) (-1.34) (-2.02) Seniority_renovation -0.00 -0.00 -0.00 -0.00* -0.00 -0.00 -0.00 -0.00 -0.00 -0.00 (-1.27) (-0.76) (-1.12) (-1.99) (-1.04) (-1.50) (-0.81) (-1.55) (-0.85) (-1.26) Structure 0.07 0.09 0.14† 0.04 0.12 0.04 0.24** 0.20* 0.14† (0.92) (1.09) (1.90) (0.60) (1.51) (0.61) (3.14) (2.45) (1.83) Height_regime_1 0.05 0.05 0.09 -0.00 0.02 (0.43) (0.39) (0.85) (-0.04) (0.21) Height_regime_2 0.02 -0.02 0.05 -0.04 0.01 (0.11) (-0.14) (0.35) (-0.24) (0.05) Height_regime_3 0.27† 0.09 0.22† -0.05 0.08 (1.86) (0.65) (1.67) (-0.34) (0.54) Height_regime_4 0.03 -0.01 0.04 -0.09 -0.04 (0.20) (-0.07) (0.36) (-0.61) (-0.34) Height_regime_5 0.09 0.09 0.09 0.04 0.03 (0.69) (0.74) (0.81) (0.33) (0.25) Height_regime_6 0.12 0.03 0.08 -0.00 0.02 (0.75) (0.18) (0.62) (-0.02) (0.17) Height_regime_7 0.24 -0.09 0.11 -0.10 0.03 (1.44) (-0.56) (0.74) (-0.58) (0.16) Height_regime_8 0.00 0.00 0.00 0.00 0.00 (.) (.) (.) (.) (.) Height_regime_9 0.19 0.18 0.19† 0.13 0.13 (1.59) (1.50) (1.76) (1.06) (1.13) Height_regime_10 0.25† 0.22 0.27* 0.16 0.19 (1.82) (1.60) (2.16) (1.12) (1.45) S_useful 0.01*** 0.03*** 0.01*** 0.03*** (9.63) (9.82) (10.24) (9.54) S_useful_sq -0.00*** -0.00*** (-6.58) (-6.15) S_constructed 0.01*** 0.03*** 0.01*** 0.02*** (8.86) (9.93) (9.37) (9.28) S_constructed_sq -0.00*** -0.00*** (-6.82) (-5.77) _cons 10.00*** 9.96*** 9.92*** 9.49*** 9.91*** 9.42*** 9.90*** 9.61*** 9.97*** 9.53*** (65.73) (53.81) (66.69) (64.32) (54.27) (51.69) (64.74) (69.58) (52.63) (50.59) Turning points 116,15 117,77 126,04 134,33 F statistic 38.18*** 22.13*** 40.08*** 49.78*** 22.84*** 28.81*** 36.84*** 50.63*** 20.78*** 25.56*** R-sq 0.68 0.72 0.69 0.75 0.73 0.78 0.67 0.72 0.71 0.76 # 173.00 173.00 173.00 173.00 173.00 173.00 173.00 183.00 173.00 173.00 Source: Authors’ computations.† p < 0.10. * p < 0.05. ** p < 0.01. *** p < 0.001 ===== PAGE 18 (score 205, numeric-paren 161, table-words 1, var-words 7) ===== 406 Zilișteanu, I. R., Muntean, R., Gherghina, Ș. C., Vintilă, G., Barbu, T. C. Table no. 12 – The output of OLS regressions regarding the impact of the entire range of housing structural attributes on the sale price (duration until transaction completion included) Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) No_rooms 0.23*** 0.25*** (9.05) (9.70) Comfort 0.34*** 0.31*** 0.26*** 0.14* 0.20** 0.12* 0.26*** 0.14* 0.21** 0.12* (5.81) (5.23) (4.32) (2.47) (3.30) (2.05) (4.26) (2.50) (3.25) (2.03) Floor -0.01* -0.01 -0.01 -0.01** -0.01 -0.01* -0.01 -0.01* -0.01 -0.01* (-2.01) (-1.47) (-1.44) (-2.64) (-1.36) (-2.41) (-1.20) (-1.98) (-1.35) (-2.18) No_bathrooms 0.08 0.03 0.07 0.05 0.04 0.03 0.09† 0.07 0.06 0.05 (1.65) (0.55) (1.48) (1.14) (0.90) (0.64) (1.79) (1.59) (1.15) (1.04) No_balconies -0.04 -0.04 -0.05 -0.03 -0.06 -0.04 -0.08† -0.06 -0.09† -0.06 (-0.82) (-0.97) (-1.12) (-0.73) (-1.39) (-0.92) (-1.78) (-1.38) (-1.92) (-1.40) No_parking 0.21** 0.23** 0.09 0.16** 0.13† 0.20** 0.06 0.16* 0.09 0.17* (3.11) (3.19) (1.35) (2.63) (1.69) (2.92) (0.81) (2.52) (1.10) (2.30) Seniority_construction -0.00*** -0.01*** -0.00 -0.00 -0.00 -0.00† -0.00 -0.00† -0.00 -0.00* (-3.87) (-4.61) (-0.43) (-1.37) (-0.88) (-1.91) (-1.01) (-1.73) (-1.32) (-2.09) Seniority_renovation -0.00 -0.00 -0.00 -0.00† -0.00 -0.00 -0.00 -0.00 -0.00 -0.00 (-1.18) (-0.70) (-1.10) (-1.89) (-1.03) (-1.43) (-0.81) (-1.40) (-0.86) (-1.19) Structure 0.06 0.08 0.14† 0.02 0.12 0.03 0.25** 0.21* 0.12 (0.72) (0.91) (1.85) (0.32) (1.48) (0.34) (3.11) (2.44) (1.57) Duration -0.00 -0.00 -0.00 -0.00 0.00 -0.00 0.00 -0.00† 0.00 -0.00 (-0.93) (-0.69) (-0.11) (-1.37) (0.05) (-1.08) (0.09) (-1.73) (0.19) (-0.97) Height_regime_1 0.06 0.05 0.11 -0.01 0.04 (0.51) (0.38) (1.00) (-0.06) (0.34) Height_regime_2 0.03 -0.02 0.06 -0.04 0.02 (0.18) (-0.15) (0.47) (-0.26) (0.15) Height_regime_3 0.27† 0.09 0.22† -0.05 0.08 (1.89) (0.65) (1.72) (-0.34) (0.56) Height_regime_4 0.04 -0.01 0.06 -0.09 -0.03 (0.28) (-0.08) (0.49) (-0.62) (-0.23) Height_regime_5 0.10 0.09 0.11 0.04 0.04 (0.78) (0.72) (0.95) (0.30) (0.37) Height_regime_6 0.12 0.03 0.09 -0.00 0.03 (0.78) (0.18) (0.67) (-0.02) (0.20) Height_regime_7 0.25 -0.09 0.13 -0.10 0.04 (1.52) (-0.56) (0.87) (-0.59) (0.27) Height_regime_8 0.00 0.00 0.00 0.00 0.00 (.) (.) (.) (.) (.) Height_regime_9 0.20 0.18 0.20† 0.13 0.14 (1.65) (1.48) (1.87) (1.03) (1.22) Height_regime_10 0.26† 0.22 0.28* 0.16 0.20 (1.86) (1.59) (2.25) (1.10) (1.52) S_useful 0.01*** 0.03*** 0.01*** 0.03*** (9.42) (9.89) (9.96) (9.53) S_useful_sq -0.00*** -0.00*** (-6.74) (-6.25) S_constructed 0.01*** 0.03*** 0.01*** 0.03*** (8.65) (10.13) (9.09) (9.20) S_constructed_sq -0.00*** -0.00*** (-7.03) (-5.84) _cons 10.03*** 9.97*** 9.92*** 9.50*** 9.91*** 9.42*** 9.90*** 9.61*** 9.97*** 9.53*** (64.51) (53.63) (65.47) (64.39) (53.99) (51.67) (63.55) (69.98) (52.33) (50.55) Turning points 115,63 117,20 125,93 133,54 F statistic 34.42*** 20.92*** 35.86*** 45.67*** 21.50*** 27.45*** 32.95*** 46.40*** 19.56*** 24.32*** R-sq 0.68 0.72 0.69 0.76 0.73 0.78 0.67 0.73 0.71 0.76 # 173.00 173.00 173.00 173.00 173.00 173.00 173.00 183.00 173.00 173.00 Source: Authors’ computations.† p < 0.10. * p < 0.05. ** p < 0.01. *** p < 0.001 ===== PAGE 19 (score 132, numeric-paren 94, table-words 2, var-words 12) ===== Scientific Annals of Economics and Business, 2019, Volume 66, Issue 3, pp. 389-413 407 Further, when adding the squared term of the useful surface (Eq. 4 and Eq. 6) and constructed surface (Eq. 8 and Eq. 10), we notice a concave association between the latter and the price of the apartment, similar to Ooi et al. (2014), Brecard et al. (2018). Consequently, the threshold for useful surface is located between 115-117 square meters and between 125- 134 squared meter in case of constructed surface. 4.3 Robustness checks A broad description of various distributions may be gathered via quantile regression which considers the overall conditional distribution, not just the conditional mean effect (Wen et al., 2019).As different to linear regression, which estimates a conditional mean function, quantile regression estimates a conditional quantile function, in which a quantile of the response variable’s conditional distribution is showed as a function of covariates (Liao and Wang, 2012). As well, quantile regression estimates are more informative and robust to outliers (Zhang and Yi, 2017). The influence of the entire range of housing structural attributes on the sale price, except constructed surface, estimated via quantile regressions is exhibited in Table no. 13. Table no. 13 – The outcomes of quantile regressions towards the impact of the entire range of housing structural attributes (except constructed surface) on the sale price Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) Quantile .1 .2 .3 .4 .5 .6 .7 .8 .9 S_useful 0.01*** 0.01*** 0.01*** 0.01*** 0.01*** 0.01*** 0.01*** 0.01*** 0.01*** (6.85) (6.34) (9.78) (14.08) (14.41) (9.20) (6.82) (6.47) (5.62) Comfort 0.14 0.12 0.18** 0.18*** 0.20*** 0.24*** 0.26** 0.29** 0.30* (1.51) (1.26) (2.73) (3.91) (4.09) (3.86) (3.25) (2.91) (2.57) Floor 0.00 -0.00 -0.01 -0.00 -0.01 -0.00 -0.01 -0.01 -0.02 (0.30) (-0.42) (-0.95) (-0.65) (-1.30) (-0.55) (-1.04) (-1.52) (-1.60) No_bathrooms 0.03 0.10 0.07 0.07* 0.05 0.12* 0.13* 0.07 0.07 (0.35) (1.25) (1.31) (2.07) (1.30) (2.32) (2.10) (0.96) (0.72) No_balconies 0.01 0.00 0.01 -0.02 -0.02 -0.04 -0.05 -0.06 -0.02 (0.11) (0.05) (0.12) (-0.69) (-0.59) (-0.95) (-0.92) (-0.88) (-0.23) No_parking -0.07 0.06 0.04 0.03 0.02 0.10 0.09 0.06 0.10 (-0.59) (0.50) (0.53) (0.65) (0.27) (1.37) (0.92) (0.51) (0.76) Seniority_construction 0.00 -0.00 0.00 -0.00 -0.00 0.00 -0.00 0.00 -0.00 (0.13) (-0.45) (0.02) (-0.46) (-0.44) (0.16) (-0.37) (0.53) (-0.11) Seniority_renovation -0.00 -0.00 -0.00* -0.01** -0.00* -0.00† -0.00 -0.00 0.00 (-0.96) (-1.27) (-1.98) (-3.34) (-2.44) (-1.91) (-1.00) (-1.05) (0.13) Structure 0.87*** -0.02 0.06 0.04 0.07 0.14† 0.08 -0.09 -0.02 (7.22) (-0.16) (0.76) (0.74) (1.12) (1.78) (0.82) (-0.75) (-0.16) Duration -0.00 -0.00 -0.00 -0.00 -0.00 -0.00 -0.00 -0.00 -0.00 (-1.05) (-0.34) (-0.02) (-0.24) (-0.86) (-0.06) (-0.13) (-0.62) (-0.61) _cons 9.09*** 9.97*** 9.90*** 9.97*** 9.97*** 9.83*** 9.98*** 10.20*** 10.15*** (37.81) (40.26) (60.83) (86.47) (82.78) (62.00) (49.16) (41.19) (34.79) Pseudo R-sq 0.4887 0.5165 0.5270 0.5213 0.5132 0.5084 0.4991 0.4787 0.4593 # 173.00 173.00 173.00 173.00 173.00 173.00 173.00 173.00 173.00 Source: Authors’ computations.† p < 0.10. * p < 0.05. ** p < 0.01. *** p < 0.001 Thus, the outcomes provide support for a positive influence of useful area (Eq. 1 – Eq. 9) as in Liao and Wang (2012), McCord et al. (2018), comfort (Eq. 3 – Eq. 9), number of bathrooms (Eq. 4, Eq. 6 and, Eq. 7) similar Zhang and Leonard (2014), structure (Eq. 1 and ===== PAGE 20 (score 153, numeric-paren 108, table-words 5, var-words 11) ===== 408 Zilișteanu, I. R., Muntean, R., Gherghina, Ș. C., Vintilă, G., Barbu, T. C. Eq. 6). The seniority from renovation reveals a negative influence, as opposed to Zhang and Leonard (2014). In addition, the floor level does not reveal a statistically significant relation with sale price, on the contrary to Liao and Wang (2012). Besides, opposite to Wen et al. (2019), the number of years since construction is not statistically significant. Table no. 14 shows the quantile regression outcomes when adding the squared term of useful surface. Therefore, similar to Kim et al., 2015, the non-linear relationship between surface and sale price is confirmed. The identified thresholds are decreasing gradually from the first to the last quantile, being located between 104 and 117 square meters. Table no. 14 – The outcomes of quadratic quantiles regressions towards the impact of the entire range of housing structural attributes (except constructed surface) on the sale price Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) Quantile .1 .2 .3 .4 .5 .6 .7 .8 .9 S_useful 0.03*** 0.03*** 0.03*** 0.03*** 0.03*** 0.03*** 0.03*** 0.04*** 0.04*** (5.56) (9.94) (10.65) (10.89) (8.98) (8.31) (6.46) (6.84) (7.50) S_useful_sq -0.00*** -0.00*** -0.00*** -0.00*** -0.00*** -0.00*** -0.00*** -0.00*** -0.00*** (-3.73) (-6.99) (-7.35) (-7.54) (-6.15) (-5.83) (-4.57) (-4.92) (-5.69) Comfort 0.11 0.10† 0.11* 0.10* 0.13* 0.16* 0.18* 0.16 0.03 (1.14) (1.82) (2.15) (1.98) (2.27) (2.46) (2.15) (1.65) (0.24) Floor -0.01 -0.01* -0.01** -0.01 -0.00 -0.00 -0.01 -0.02† -0.03** (-1.24) (-2.02) (-2.73) (-1.50) (-0.15) (-0.68) (-1.37) (-1.88) (-3.32) No_bathrooms 0.08 0.08* 0.08* 0.08* 0.10* 0.12* 0.11† 0.03 0.04 (1.17) (2.09) (2.19) (2.13) (2.38) (2.47) (1.75) (0.48) (0.47) No_balconies -0.05 -0.00 -0.02 -0.01 -0.01 -0.02 -0.03 -0.05 -0.02 (-0.86) (-0.02) (-0.54) (-0.20) (-0.37) (-0.56) (-0.45) (-0.72) (-0.30) No_parking 0.11 0.23*** 0.18*** 0.25*** 0.23*** 0.22** 0.29** 0.27* 0.12 (1.05) (3.90) (3.38) (4.65) (3.59) (3.01) (3.20) (2.56) (1.00) Seniority_construction -0.00 -0.00† -0.00† -0.00† -0.00 -0.00 0.00 0.00 0.00 (-1.00) (-1.87) (-1.74) (-1.73) (-0.60) (-0.19) (0.43) (1.29) (0.26) Seniority_renovation -0.01† -0.01** -0.01** -0.00** -0.00* -0.00† -0.00 -0.00 -0.00 (-1.83) (-2.97) (-3.05) (-2.93) (-2.25) (-1.66) (-0.94) (-1.36) (-0.84) Structure -0.04 -0.01 0.02 0.04 0.09 0.06 0.06 -0.10 -0.07 (-0.33) (-0.21) (0.29) (0.60) (1.19) (0.81) (0.58) (-0.86) (-0.55) Duration -0.00 -0.00 -0.00 -0.00 -0.00 -0.00 -0.00 -0.00 -0.00† (-1.46) (-0.94) (-0.82) (-0.70) (-0.09) (-0.59) (-0.87) (-1.39) (-1.73) _cons 9.58*** 9.38*** 9.45*** 9.34*** 9.26*** 9.27*** 9.25*** 9.36*** 9.52*** (39.52) (66.34) (73.17) (72.48) (60.77) (54.64) (42.84) (37.43) (34.37) Turning points 117.44 112.23 114.21 113.81 115.25 112.43 111.68 109.68 104.02 Pseudo R-sq 0.5844 0.6082 0.5934 0.5710 0.5546 0.5453 0.5363 0.5186 0.5181 # 173.00 173.00 173.00 173.00 173.00 173.00 173.00 173.00 173.00 Source: Authors’ computations.† p < 0.10. * p < 0.05. ** p < 0.01. *** p < 0.001 The impact of the entire range of housing structural attributes on the sale price, except useful surface, assessed via quantile regressions is exhibited in Table no. 15. The outcomes confirms the positive influence of constructed surface (Eq. 1 – Eq. 9), comfort (Eq. 2 – Eq. 9), the number of bathrooms (Eq. 3, Eq. 6 and, Eq. 7) and structure (Eq. 1 – Eq. 5) on sale price. The number of balconies (Eq. 1, Eq. 5 and, Eq. 6) shows a negative, but weak, negative impact on dwellings’ price. In case of the number of parking (Eq. 1 and Eq. 5) the influence is mixed, but its statistical significance is weak. Table no. 16 reveals the output of quantile regressions when adding the squared term of constructed surface. Thereby, kindred with the outcomes from Table no. 14, the non-linear association is consolidated. The computed thresholds are located between 117 and 145 square meters. ===== PAGE 21 (score 269, numeric-paren 210, table-words 2, var-words 9) ===== Scientific Annals of Economics and Business, 2019, Volume 66, Issue 3, pp. 389-413 409 Table no. 15 – The outcomes of quantile regressions towards the impact of the entire range of housing structural attributes (except useful surface) on the sale price Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) Quantile .1 .2 .3 .4 .5 .6 .7 .8 .9 S_constructed 0.01*** 0.01*** 0.01*** 0.01*** 0.01*** 0.01*** 0.01*** 0.01*** 0.01*** (6.74) (6.59) (8.25) (9.41) (9.90) (8.45) (7.48) (6.12) (3.84) Comfort 0.14 0.16† 0.20** 0.20** 0.21** 0.25*** 0.25*** 0.30** 0.31* (1.42) (1.81) (2.80) (3.21) (3.25) (3.65) (3.38) (3.23) (2.08) Floor 0.00 -0.00 -0.01 -0.00 -0.00 -0.00 -0.01 -0.01 -0.02 (0.06) (-0.42) (-1.41) (-0.28) (-0.18) (-0.62) (-1.08) (-1.40) (-1.24) No_bathrooms 0.02 0.05 0.10† 0.06 0.06 0.12* 0.17** 0.11 0.11 (0.23) (0.79) (1.78) (1.33) (1.12) (2.25) (2.84) (1.44) (0.92) No_balconies -0.12† -0.02 -0.06 -0.07 -0.10* -0.08† -0.07 -0.03 -0.01 (-1.67) (-0.37) (-1.08) (-1.44) (-2.10) (-1.67) (-1.29) (-0.39) (-0.08) No_parking -0.26* -0.01 -0.09 0.03 0.13† 0.07 0.04 0.05 0.06 (-2.14) (-0.12) (-1.11) (0.38) (1.76) (0.90) (0.43) (0.42) (0.34) Seniority_construction 0.00 -0.00 -0.00 -0.00 -0.00 -0.00 -0.00 -0.00 -0.00 (0.12) (-0.15) (-0.34) (-1.00) (-1.36) (-0.59) (-0.71) (-0.31) (-0.01) Seniority_renovation -0.01 -0.00 -0.00 -0.00 -0.00 -0.00 -0.00 -0.00 0.00 (-1.42) (-1.31) (-1.20) (-1.21) (-0.89) (-1.04) (-1.04) (-1.30) (0.53) Structure 0.84*** 0.46*** 0.25** 0.25** 0.27** 0.13 0.09 0.05 0.07 (6.52) (4.17) (2.79) (3.17) (3.30) (1.52) (0.99) (0.43) (0.37) Duration -0.00 0.00 0.00 -0.00 -0.00 -0.00 -0.00 0.00 -0.00 (-0.37) (0.75) (0.31) (-0.12) (-0.44) (-0.07) (-0.06) (0.29) (-0.10) _cons 9.45*** 9.60*** 9.91*** 9.87*** 9.79*** 9.93*** 10.01*** 10.08*** 10.06*** (37.15) (43.72) (55.86) (63.36) (60.99) (57.76) (53.26) (43.33) (26.71) Pseudo R-sq 0.4729 0.4941 0.4960 0.4844 0.4753 0.4788 0.4792 0.4667 0.4491 # 173.00 173.00 173.00 173.00 173.00 173.00 173.00 173.00 173.00 Source: Authors’ computations.† p < 0.10. * p < 0.05. ** p < 0.01. *** p < 0.001 Table no. 16 – The outcomes of quadratic quantile regressions towards the impact of the entire range of housing structural attributes (except useful surface) on the sale price Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) Quantile .1 .2 .3 .4 .5 .6 .7 .8 .9 S_constructed 0.03*** 0.02*** 0.02*** 0.02*** 0.03*** 0.03*** 0.02*** 0.03*** 0.03*** (7.03) (8.59) (7.78) (8.25) (7.54) (8.96) (6.11) (7.48) (6.27) S_constructed_sq -0.00*** -0.00*** -0.00*** -0.00*** -0.00*** -0.00*** -0.00*** -0.00*** -0.00*** (-4.70) (-5.31) (-4.66) (-5.28) (-5.13) (-6.43) (-4.49) (-5.30) (-4.63) Comfort 0.04 0.09 0.14* 0.14* 0.15* 0.17** 0.21* 0.19* 0.13 (0.47) (1.59) (2.31) (2.45) (2.04) (2.78) (2.42) (2.19) (1.08) Floor -0.00 -0.01† -0.01* -0.01 0.00 -0.00 -0.01 -0.02* -0.02* (-0.34) (-1.72) (-2.09) (-1.12) (0.10) (-0.58) (-1.45) (-2.18) (-2.32) No_bathrooms 0.07 0.10* 0.09† 0.10* 0.13* 0.15** 0.14* 0.07 0.11 (1.13) (2.45) (1.97) (2.21) (2.37) (3.27) (2.09) (1.08) (1.20) No_balconies -0.12* -0.03 -0.05 -0.03 -0.08 -0.02 -0.01 -0.04 -0.06 (-2.09) (-0.86) (-1.22) (-0.71) (-1.60) (-0.55) (-0.09) (-0.69) (-0.68) No_parking 0.15 0.10 0.06 0.23*** 0.22** 0.21** 0.26** 0.17† 0.06 (1.58) (1.49) (0.84) (3.48) (2.73) (2.99) (2.68) (1.73) (0.42) Seniority_construction -0.00* -0.00 -0.00 -0.00 -0.00 -0.00 0.00 0.00 0.00 (-2.06) (-1.15) (-1.13) (-1.14) (-1.59) (-0.60) (0.03) (1.58) (0.60) Seniority_renovation -0.00† -0.00† -0.00† -0.00* -0.00 -0.00 -0.00 -0.01† -0.00 (-1.75) (-1.96) (-1.72) (-2.35) (-1.19) (-1.58) (-0.99) (-1.78) (-0.91) Structure 0.11 0.21** 0.23** 0.17* 0.05 0.06 0.04 0.03 0.04 (1.10) (2.92) (3.17) (2.44) (0.56) (0.83) (0.37) (0.31) (0.24) Duration -0.00 0.00 -0.00 -0.00 -0.00 -0.00 0.00 -0.00 -0.00 (-0.68) (0.27) (-0.01) (-0.86) (-0.73) (-0.52) (0.02) (-1.29) (-0.96) _cons 9.47*** 9.40*** 9.51*** 9.38*** 9.46*** 9.34*** 9.37*** 9.35*** 9.50*** (44.30) (63.29) (61.86) (62.97) (51.45) (58.64) (41.71) (40.68) (30.96) Turning points 130.35 141.10 145.46 136.23 127.84 121.34 118.38 122.88 117.77 Pseudo R-sq 0.5574 0.5769 0.5539 0.5284 0.5185 0.5133 0.5152 0.5028 0.5006 # 173.00 173.00 173.00 173.00 173.00 173.00 173.00 173.00 173.00 Source: Authors’ computations.† p < 0.10. * p < 0.05. ** p < 0.01. *** p < 0.001 ===== PAGE 22 (score 19, numeric-paren 0, table-words 0, var-words 18) ===== 410 Zilișteanu, I. R., Muntean, R., Gherghina, Ș. C., Vintilă, G., Barbu, T. C. 5. CONCLUDING REMARKS Current study aimed to explore the influence of the structural attributes related to an apartment on its sale price in Bucharest, Romania, over 2014-2018. Several structural attributes were selected, among which OLS regression outcomes provide support for a positive impact, respectively the number of rooms, comfort, the number of bathrooms, the number of parking, the number of balconies and, structure. On the contrary, in case of floor level and seniority from construction, the OLS estimates reveals a negative relationship. Nevertheless, we notice concave links between useful surface and sale price, as well as between constructed surface and sale price. When estimating via quantile regressions, the associations are reinforced. The main limitation of current study is driven by OLS estimation regardless of the spatial effects (Z. Z. Huang et al., 2017). The issue with traditional hedonic pricing models is the insufficiency towards depicting the fundamental link between property price and geographical spatial interaction (Hui et al., 2012). Therefore, as future research avenues, spatial econometric modeling will be considered. Acknowledgements The research is supported by RMI project „Ansamblu de indici imobiliari structurați pentru piața românească” - Proiectul este cofinanțat prin Programul Operațional Competitivitate 2014-2020, Axa Prioritară 2 „Tehnologia Informației și Comunicațiilor (TIC) pentru o economie digitală competitivă”, Obiectiv specific 2.2 „Creșterea contribuției sectorului TIC pentru competitivitatea economică”, Acțiunea 2.2.1 „Sprijinirea creșterii valorii adăugate generate de sectorul TIC și a inovării în domeniu prin dezvoltarea de clustere”, Prioritatea de Investiții „PI2.2 Dezvoltarea produselor și serviciilor, a comerțului electronic, a comerțului electronic și a cererii de TIC” - cod mySMIS 115645. References Baranzini, A., and Schaerer, C., 2011. A sight for sore eyes: Assessing the value of view and land use in the housing market. Journal of Housing Economics, 20(3), 191-199. http://dx.doi.org/10.1016/j.jhe.2011.06.001 Benson, E. D., Hansen, J. L., Schwartz, A. L., and Smersh, G. T., 1998. Pricing residential amenities: The value of a view. The Journal of Real Estate Finance and Economics, 16(1), 55-73. http://dx.doi.org/10.1023/A:1007785315925 Brecard, D., Boennec, R. L., and Salladarre, F., 2018. Accessibility, local pollution and housing prices. Evidence fromNantes Metropole, France. Economie et Statistique. Economie & Statistique(500- 502), 97-115. http://dx.doi.org/10.24187/ecostat.2018.500t.1952 Camins-Esakov, J., and Vandegrift, D., 2018. Impact of a light rail extension on residential property values. Research in Transportation Economics, 67, 11-18. http://dx.doi.org/10.1016/j.retrec.2017.04.004 Chen, W. Y., and Jim, C. Y., 2010. Amenities and disamenities: A hedonic analysis of the heterogeneous urban landscape in Shenzhen (China). The Geographical Journal, 176, 227-240. http://dx.doi.org/10.1111/j.1475-4959.2010.00358.x Conroy, S., and Milosch, J. L., 2011. An Estimation of the Coastal Premium for Residential Housing Prices in San Diego County. The Journal of Real Estate Finance and Economics, 42(2), 211-228. http://dx.doi.org/10.1007/s11146-009-9195-x