Portal de Periódicos da CAPES

The Impact of Industrial Sites on Residential Property Values: A Hedonic Pricing Analysis from the Netherlands

2010; Routledge; Volume: 45; Issue: 5 Linguagem: Inglês

10.1080/00343401003601925

1360-0591

Friso de Vor, H.L.F. de Groot,

Regional Economics and Spatial Analysis

Abstract De Vor F. and De Groot H. L. F. The impact of industrial sites on residential property values: a hedonic pricing analysis from the Netherlands, Regional Studies. Industrial sites cause several negative externalities. In order to quantify these negative effects, the impact of distance to industrial sites on residential property values is estimated. Data on houses sold in the Randstad region and in the Province of North-Brabant (both located in the Netherlands) in the year 2005 are used, together with data on the characteristics of industrial sites in the same regions and period. The distance to an industrial site exhibits a statistically significant negative effect on the value of residential properties. However, the effect is largely localized within a relatively short distance from the nearest industrial site. De Vor F. et De Groot H. L. F. L'impact des sites industriels sur la valeur des logements: une analyse néerlandaise de la détermination hédonique des prix, Regional Studies. Les sites industriels ont plusieurs effets externes négatifs. Pour quantifier ces effets négatifs, on estime l'impact de la distance aux sites industriels sur la valeur des logements. On emploie des données sur les logements vendus dans la région Randstad et dans la province de Brabant du Nord (les deux étant situées en Hollande) pour l'année 2005, conjointement avec des données sur les caractéristiques des sites industriels situés dans les mêmes régions à la même époque. Il s'avère que la distance à un site industriel a un effet négatif statistiquement important sur la valeur des logements. Toujours est-il que l'effet est dans une large restreint à une distance relativement courte du site industriel le plus proche. Sites industriels Effets externes négatifs Détermination hédonique des prix De Vor F. und De Groot H. L. F. Auswirkung von Industriestandorten auf den Wert von Wohnimmobilien: eine hedonische Preisanalyse in den Niederlanden, Regional Studies. Industriestandorte verursachen mehrere negative Externalitäten. Zur Quantifizierung dieser negativen Effekte wird die Auswirkung der Entfernung von Industriestandorten auf den Wert von Wohnimmobilien geschätzt. Hierfür kommen die Daten von Häusern zum Einsatz, die im Jahr 2005 in der Randstad-Region sowie in der Provinz Nordbrabant (beide in den Niederlanden) verkauft wurden, sowie die Daten über die Merkmale von Industriestandorten in denselben Regionen und im selben Zeitraum. Die Entfernung von einem Industriestandort wirkt sich auf statistisch signifikante Weise negativ auf den Wert von Wohnimmobilien aus. Allerdings ist diese Auswirkung größtenteils auf eine relativ kurze Entfernung bis zum nächsten Industriestandort begrenzt. Industriestandorte Negative Externalitäten Hedonische Preiskalkulation De Vor F. y De Groot H. L. F. El efecto de los polígonos industriales en los valores de bienes inmuebles residenciales: análisis de precios hedónicos en los Países Bajos, Regional Studies. Los polígonos industriales causan varios efectos externos negativos. A fin de cuantificar estos efectos negativos, calculamos el impacto de la distancia a los polígonos industriales en los valores de bienes inmuebles residenciales. Utilizamos datos de viviendas vendidas en la región de Randstad y la provincia de Brabante Septentrional (ambas en los Países Bajos) en el año 2005, junto con datos sobre las características de los polígonos industriales en los mismos periodos y regiones. La distancia a un polígono industrial muestra un efecto negativo estadísticamente significativo sobre el valor de los bienes inmuebles residenciales. Sin embargo, el efecto está en gran medida localizado dentro de una distancia relativamente corta desde el polígono industrial más cercano. Polígonos industriales Efectos externos negativos Precios hedónicos Keywords: Industrial sitesNegative externalitiesHedonic pricingJEL classifications: O18R30R52 Acknowledgements This research was funded by the BSIK programme ‘Vernieuwend Ruimtegebruik’. The authors gratefully acknowledge the Dutch Association of Real Estate Agents (NVM) for providing the data on residential property values. They are grateful to Erhan Demirel, Laura de Dominicis, Erik Louw, Ioulia Ossokina, Cees-Jan Pen, Piet Rietveld, Willemijn van der Straaten, and two anonymous reviewers for useful comments. The authors also would like to thank Chris Jacobs for research assistance. The usual disclaimer applies. Notes Five categories of industrial sites have been defined: ‘heavy industry’, ‘sea harbour’, ‘miscellaneous’, ‘high-tech’, and ‘transport’. This typology of industrial sites, as defined by IBIS, is associated with the nature of the industrial site concerned. See, for example, Zeiss Citation(1999), who analyses the results of sixty-nine studies dealing with the negative effects of industrial facilities on their direct vicinity. In general, an industrial site can be considered as a collective location for the establishment of firms (Bak, Citation1985). Sites designated exclusively for offices are not covered by this definition. The Randstad region comprises the following areas: the Province of North-Holland (excluding Alkmaar and the surrounding area and the northern part of North-Holland) and the Provinces of South-Holland and Utrecht. For more details about regional definitions, see Rietveld and Wagtendonk Citation(2004). The examined area originally comprised 1718 industrial sites. Information on the ‘type of industrial site’ is only available for 1201 sites. Those houses closest to an industrial site but for which no information on the type of industrial site was available were eliminated from the sample. An industrial site is defined as ‘a location which the land-use plans deem suitable for activities in the branches of commerce, manufacturing, commercial services, industry, and whose original design is larger than one hectare’. Heavy-industry sites are industrial sites on which it is intended to establish highly undesirable facilities in terms of environmental nuisance (for example, landfills, demolition dumps, waste sites, hazardous manufacturing facilities, etc.). Sea harbours comprise large-scale sites featuring freight facilities for maritime shipping activities. Miscellaneous sites are commonly seen as regular industrial sites that include a large variety of activities (ranging from small to medium polluting activities), on the condition that these sites may not be categorized as high-tech or transport. High-tech sites are those developed for high-tech production activities and research and development activities. Transport is characterized by a dominant presence of logistics and wholesale activities. NVM real-estate agents deal with approximately 70% of all property transactions in the Netherlands. These data sets were provided by SPINLab, VU University Amsterdam. Rents are controlled in the Netherlands, implying that their values are determined by a restricted system of points that highly disregards neighbourhood characteristics. The authors, therefore, do not take into account rented housing in the analysis. Gross area is defined as the total area designated as the industrial site, including the area used for infrastructure, water, green space, and other forms of public open space. Details are available from the authors on request. The results might be biased due to a possibly negative correlation between distance to the industrial sites and distance to, for example, public goods provided in the city centre. This problem becomes more severe if many industrial sites are located in the urban fringe. This bias is likely to be limited in the present case because Equation(1) the sample comprises both inner-city industrial sites and fringe sites (for example, greenfield sites); and Equation(2) the empirical analysis controls for a range of characteristics of city centres that may positively affect house prices, thus reducing the bias due to an omitted variable (namely, distance to ‘goods’ typically provided in city centres such as railway stations and employment). Concerning the impact of the age of the dwelling on house prices, alternative models have been estimated by applying a linear and a quadratic specification for the year of the transaction. The estimated coefficients are consistent with the general findings regarding sign as well as statistical significance of the estimated effects. Details are available from the authors on request. The representative property is defined as the hypothetical property for which all explanatory variables (namely, attributes of the property), except for distance to the industrial site, are set at their respective means. The adjusted R 2 of both models is 0.774. The Akaike Information Criterion (AIC) for the respective models is −0.251 for the piecewise model, and −0.250 for the logistic model. The model with a lower AIC is typically preferred. The estimation results of the other models that are related to the interaction between site size and distance are available from the authors on request. Sampling along the spatial dimensions concerned creates data sets consisting of the following number of observations: fewer than 500, N = 27 711; fewer than 1000, N = 50 660; fewer than 1500, N = 62 485; and fewer than 2000, N = 67 600. The estimation results related to this sensitivity analysis are available from the authors on request. Sampling along the regional dimensions concerned creates data sets consisting of the following number of observations: Randstad, N = 51 689; and North-Brabant, N = 18 995. This has been confirmed by running an alternative regression (ordinary least-squares – OLS) using the natural logarithm of distance. The regression yields statistically significant estimates of the log of distance (at the 1% significance level). Estimation results are available from the authors on request. The additional models comprise four basic regression runs per region for different compositions of the samples: a sample with only small sites (fewer than 35 ha), a sample with only large sites (greater than 35 ha), and samples with only heavy and non-heavy sites, respectively. The estimation results are available from the authors on request. The estimation results are available from the authors on request.