1# RECORD 96e0ec640b350c1ec477b4a65c34f99c77d535032# TITLE: Identifying Determinants of Residential Property Values in South London3# YEAR: 2011.0 VENUE: nan DOI: 10.2478/v10135-011-0004-04# AUTHORS: D. E. May; Adam R. Corbin; Philip D. Hollins5# PAGES: 10 CANDIDATE TABLE PAGES (1-based): [5]6# ABSTRACT: Identifying Determinants of Residential Property Values in South London This paper advances the academic work conducted on identifying some of the determinants of residential property values in different countries. Such determinants might include positive factors such as access to amenities, or negative factors such as high voltage overhead transmission lines (HVOTL). The objective explored here is to extend this research in the UK by considering the particular case of residential property values in South London. A number of determinants is considered, grouped and a model which explains the effect of the determinants on value is produced. The elasticities between property values and the distances from different determinants were also analysed to help highlight relationships between the determinants and residential property values. The results showed a relationship between residential property values and the determinants, and a relationship between the determinants themselves.789===== PAGE 1 (score 5, numeric-paren 0, table-words 0, var-words 5) =====1011 12NÁRODOHOSPODÁŘSKÝ OBZOR – REVIEW OF ECONOMIC PERSPECTIVES, 13VOL. 11, ISSUE 1, 2011, pp. 3–11, DOI: 10.2478/v10135-011-0004-0 14 15IDENTIFYING DETERMINANTS OF RESIDENTIAL PROPERTY VALUES IN 16SOUTH LONDON 17Daniel E. May, Adam R. Corbin and Philip D. Hollins 18Harper Adams University College, Newport, Shropshire, TF10 8NB, United Kingdom, 19dmay@harper-adams.ac.uk, adam.corbin@northumbria.ac.uk, phollins@harper-20adams.ac.uk 21Abstract: This paper advances the academic work conducted on identifying some of the 22determinants of residential property values in different countries. Such determinants 23might include positive factors such as access to amenities, or negative factors such as 24high voltage overhead transmission lines (HVOTL). The objective explored here is to 25extend this research in the UK by considering the particular case of residential property 26values in South London. A number of determinants is considered, grouped and a model 27which explains the effect of the determinants on value is produced. The elasticities 28between property values and the distances from different determinants were also 29analysed to help highlight relationships between the determinants and residential 30property values. The results showed a relationship between residential property values 31and the determinants, and a relationship between the determinants themselves. 32Key words: House characteristics, amenity places, house value, modelling residential 33house price, cross elasticities 34JEL Classification: R21, R32 35DOI: 10.2478/v10135-011-0004-0 36 373839===== PAGE 5 (score 34, numeric-paren 5, table-words 9, var-words 6) =====4041REVIEW OF ECONOMIC PERSPECTIVES 42 43 44 45646whether the multicollinearity problem was avoided using this strategy, collinearity was 47assessed using auxiliary regressions. According to Hill et al. (2008), detrimental 48collinearity is present when the coefficients of determination of these regressions are 49equal to or higher than 0.80. Because the coefficients of determination of the auxiliary 50regressions estimated in this research were lower than 0.80, it was concluded that 51detrimental collinearity was not present in the model. The auxiliary regressions are 52shown in Appendix B. Finally, The Goldfeld-Quandt test was used to test 53heteroscedasticity. For this purpose, the n observations in the sample were ordered by 54the magnitude of P and separate regressions were run for the first 469 and the last 469 55observations. The residual sum of squares in the two sub-regressions was used to obtain 56the ratio test RSS2/RSS1 (i.e. the Goldfeld-Quandt test) where RSS1 and RSS2 are the 57residual sum of squares of the first and second sub-regression, respectively (see 58Dougherty, 2007, p. 229). The values of RSS1 and RSS2 obtained from the sub-59regressions were 5.23 and 6.35, respectively. Using this test, heteroscedasticity was 60rejected at the 10% of significant level, but not at the 5%. As a consequence, the results 61obtained in this investigation have to be considered with caution. 62The estimation of the model described in Equation 1 is presented in Table 1, and the 63relevant elasticities are presented in Table 2. 64Table 1: Hedonic multiple regression estimation 65Variable 66Coefficient 67t-statistic 68p-value 69Intercept 70Terraced 71Semi-detached 72Leasehold 73Flat 74LnDCL 75LnDPK 76LnDRA 77LnDWS 78LnDRA*LnDPY 79LnDRA*LnDPK 80LnDWS*LnDPY 81LnDPY*LnDPK 829.54 83-0.12 84-0.08 85-0.19 86-0.33 870.03 881.08 890.87 90-0.89 91-0.05 92-0.09 930.11 94-0.08 9512.26 96-4.63 97-3.48 98-2.22 99-373 1003.63 1015.98 1025.53 103-7.77 104-5.37 105-4.99 1066.19 107-6.10 1080.000 1090.000 1100.005 1110.027 1120.002 1130.003 1140.000 1150.000 1160.000 1170.000 1180.000 1190.000 1200.000 121Adjusted R2 0.44 122S.E. of regression 0.17 123 124 125 126Table 2: Elasticities of house’s value and distance variables 127Elasticity 128Value 129∂LnP/∂LnDCL = 1300.03 131∂LnP/∂LnDPK = 1321.08 – 0.09*LnDRA – 0.08*LnDPY 133∂LnP/∂LnDRA = 134– 0.05*LnDPY – 0.09*LnDPK 135∂LnP/∂LnDWS = 136– 0.89 + 0.11*LnDPY 137 138139140===== PAGE 6 (score 16, numeric-paren 0, table-words 4, var-words 8) =====141142Volume 11, Issue 1, 2011 143 144 145 1467147The model presented in Table 1 reveals that property values are affected by different 148house characteristics. Leasehold was shown to have a negative effect on value, however 149this result has to be considered with caution as this indicator needs to be contextualised 150with the other characteristics and the data considered by the model. For example, it is 151possible to demonstrate that by keeping all the distance variables constant, a freehold 152flat is less expensive than leasehold semi-detached house. Whilst it is common sense 153that most flats are worth less than houses of any type; a freehold flat is most unusual, as 154it would have to be in contact with the ground to be registered as freehold. Therefore, a 155clear picture of the way in which house characteristics affect property values depends 156upon how these characteristics are combined, and the legal realities of home ownership 157in England and Wales. 158Table 2, on the other hand, shows that house values increases by 0.03% when the 159distance to the centre line of HVOLT (DCL) increases by 1%. This result is consistent 160with that obtained by Sims et al. (2009) and Sims and Dent (2005). That is, this 161evidence supports the existence of a relationship between value diminution and the 162presence of an HVOTL near residential properties. Table 2 also shows that the 163relationship between proximity to public parks (DPK) and property values is relative as 164it depends on the interaction between proximity to public parks, railway lines (DRA) 165and pylons (DPY). For example, the elasticity between house values and DPK is equal 166to -0.16 for houses located 1,500 metres from railway lines and pylons. For this set of 167houses, proximity to public parks increases the value of residential properties. In 168contrast, the elasticity for houses located 500 metres from these infrastructures is 0.02 169indicating that proximity to public parks decreases the property value for this set of 170houses. It could be inferred from this result that the beneficial effects of proximity to a 171public park disappears in the presence of railway lines and pylons. This finding is 172consistent with that obtained by Giaccaria and Frontuto (2007). In synthesis, residents in 173South London did not consider public parks as an isolate amenity attribute when 174purchasing residential properties but as an element within a more complex system that 175involves a number of interactions. A similar result has been observed in the case of 176proximity to railway lines; here the way in which they affect residential property values 177depends upon the interaction between the lines and the proximity to pylons and public 178parks. Finally, Table 2 shows that an interaction exists between proximity to Welling 179station (DWS) and pylons. For example, the elasticity between house values and DWS is 180equal to 0.01 for houses located 3,500 metres from pylons. In contrast, this elasticity is 181equal to -0.21 for houses located 500 metres from the pylons. This finding suggests that 182any loss of value caused by the presence of pylons is partially offset when a transport 183service is available. 184Conclusions 185The present investigation used a hedonic model to identify the effect of a number of 186determinants on residential property values in South London. The results revealed that 187these values were negatively affected by closeness to HVOTL and by house attributes 188such as where they were part of a terrace. In contrast, detached houses had higher value. 189The existence of an interaction was also found between some determinants, which 190resulted in an ambiguous effect upon house values. For example, proximity to public 191parks had only a positive effect on property values for houses located far away from 192