Back to search

Panteion University of Social and Political Sciences

Διερεύνηση των αξιών ακινήτων στη μητροπολιτική περιοχή της Αθήνας: εφαρμογή σύγχρονων μεθόδων χωρικής ανάλυσης

Abstract

dc:description

The aim of this study is the analysis of the prices of real estate in the metropolitan area of Athens and the definition of submarkets in the real estate market under study. This aim is approached with the use of the theory of hedonic prices, spatial analysis methods and spatial econometrics.In the first part of the dissertation, the general theory of hedonic prices is presented, which gives the theoretical background to study the way real estate prices are defined. In the real estate market, demand and, consequently, prices, do not behave homogenously in every type of property, geographical area, or submarket. For this reason, I present the term of submarkets and the importance of their definition, for the estimation of hedonic models. In the next part, there is an analytical presentation of methodological issues regarding the effect of the spatial structure, since the phenomenon studied in this dissertation is an economical one, evolving in space. The measurements of the prices and the property characteristics are defined in a particular spatial structure. The problem that we address is how to remove the effect of the spatial structure in order to examine the economic phenomenon of our interest, which is the real estate prices and the factors defining them. Therefore, we analyze the spatial effects that lead the systematic change of the prices and the real estate characteristics. Furthermore, spatial econometric models are discussed and their estimation methods are presented, with a special focus on spatial weight matrices, with which the spatial structure can be described. Last, in the context of the theoretical and methodological part of the dissertation, Cluster Analysis is presented, according to Silverman's approach, on which the algorithm that was used to define the submarkets in the study area was based.The analysis is based on 32992 real estate from web classified ads, the collection of which was realized on September 2013. The spatial models used for the econometric analysis and the inclusion of spatial effects are the spatial autoregressive model (SAR), the spatial error model (SEM), the spatial general model (SAC) and the spatial Durbin model (SDM). 13 ways of spatial weight matrices construction are studied, based on distance and neighborhood. The 13 ways are applied on 4 models of spatial econometrics and 52 specializations are realized. The statistical criteria used for the control of the models and the comparison of the specializations are the model fit, the variance of the residuals, the log likelihood, the Akaike information criterion and Moran's test. According to the results, the SAC with 15-nearest real estate units is the most suitable hedonic model. The SAC model was tested regarding its predictive accuracy and the first derivatives were calculated in order to measure the effect of each variable on the definition of the marginal implicit price. The results of the non-spatial OLS were used to define the specification of the hedonic prices model of the Athens real estate properties. Then we apply statistical tests for the normality of the disturbance term, for the model specification and for spatial effects, as well. Then, I compare the non-spatial OLS results, with those of SAC. The comparison was done regarding the predictive accuracy, the values of the factors and the values of standard errors. According to the results, the predictions of SAC are more accurate than the OLS ones. Also, the MWTP differ significantly between the two models and in particular, OLS mainly overestimates the effects of the variables. It is important that the standard errors are significantly reduced in the spatial model. As a result, in the SAC model the confidence interval are smaller and this way there is higher accuracy in the estimation of the model factors. As a conclusion, according to the findings of this dissertation, there are 11 factors that form the real estate prices and they contribute significantly to their spatial variation. These factors are: the size, the air-condition availability, the first, second, third, fourth and fifth or higher floor, the sea view, the age, being in the basement and the lack of a lift. Also, the results have shown that the most significant differences between prices that are estimated by the model and the prices in the ads, are found in the most expensive areas of Athens and also that the extreme estimation are noticed in real estate on the edge of the study area. Regarding the model of spatial concentration of the predicted values, this is different in the study area, according to the results obtained by spatial interpolation, the Athens area is divided in 4 submarkets. Finally, the methodology that was developed for the definitions of submarkets according to the value of relative frequency of the real estate gave 4 submarkets of the metropolitan Athens area. Though, the statistical criteria used have shown that the most suitable number of submarkets is 180. Furthermore, according to the results, the borders of the 180 neighborhoods, approach very well the areas where a difference of the spatial model is observed.

Degree

thesis:*
Grantor dc:publisher
Panteion University of Social and Political Sciences
Year dc:date
2016

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Marianthi, Stamou
  • Στάμου, Μαριάνθη

Subjects

dc:subject × 12

Rights

Language dc:language
gre

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/37967

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Marianthi, Stamou; Στάμου, Μαριάνθη. Διερεύνηση των αξιών ακινήτων στη μητροπολιτική περιοχή της Αθήνας: εφαρμογή σύγχρονων μεθόδων χωρικής ανάλυσης. Panteion University of Social and Political Sciences, 2016. http://hdl.handle.net/10442/hedi/37967