THE HEDONIC PRICING MODEL OF THE KRASNOYARSK PRIMARY REAL ESTATE MARKET

THE HEDONIC PRICING MODEL OF THE KRASNOYARSK PRIMARY REAL ESTATE MARKET

Dmitrii Vershinin

master’s student, Siberian Federal University,

Russia, Krasnoyarsk

ABSTRACT

The article discusses hedonic pricing model of the primary real estate market on the example of the city of Krasnoyarsk. The aim of the study is to build multiple regression models based on the internal and external factors of the price formation process for each city cluster. This research is relevant today, since such studies have not been conducted for the city of Krasnoyarsk in terms of identifying price formation factors of the primary real estate with help of the digital technologies.

Keywords: hedonic pricing, real estate market, regression modeling, clustering, price formation factors.

1. Introduction

Krasnoyarsk is the capital of the Siberian region called Krasnoyarsk Krai, a part of investment project “Yenisei Siberia”. Thus, this paper makes an important contribution to regional development including the understanding of city development strategy. This research paper is based on only primary sales of real estate because it shows the current situation.

The hedonic pricing is one of the preferred methods of the real estate market research. Hedonic pricing is a model that takes into account pricing factors, based on the assumptions that the price is estimated by internal and external factors [1]. The main idea of the hedonic price modeling is that each characteristic specific of an object or area contributes to the total value of real estate. In case of the price analysis of the primary real estate, the main theory of the hedonic pricing method uses different characteristics of the apartment such as physical or apartment structure characteristics (square, floor, rooms, material, age building, etc.), spatial characteristics (transport availability, social infrastructure availability, city center availability, etc.) and neighbourhood characteristics (view from the window, street-level noise or highway traffic noise, ecology, proximity to the industrial area, etc.).

The paper structure is as follows. In the beginning, a description is given of the cluster analysis results and price formation factors used to build the regression model. Then, a description of main results.

2. Raw data and cluster analysis results

It’s been collected more than 9.3 thousand primary sales available on January 2022. Data base was formed with usage of price parsing and data crawling with Screaming Frog SEO Spider. Data sources include different Russian real-estate agencies [2; 3; 4].

The following price formation factors were used for cluster analysis and building multiple regressions models. You can see them in the regression function below:

                                                                          (3)

where:

 – price per the square meter in rubles;

 – floor apartment;

 – rooms;

 – square apartment;

 – apartment material;

 – infrastructure availability index (including medium distance to the nearest school, hospital, shopping mall and city center distance);

 – social infrastructure level within 1 km radius (including public transport availability, quantity of the supermarkets, quantity of the schools, quantity of the kindergartens, quantity of the hospitals and gyms).

The cluster analysis is proposed using a machine learning software called STATISTICA. The results are shown in Table 1. The 57 residential complexes are clustered into 7 groups. In order to see cluster distribution of each group more clearly, the visualization of them are figured out and shown in Figure 2.

Table 1.

Cluster Analysis

Figure 2. Cluster analysis results

3. Results of the multiple regression modeling

You can see the results of the regression modeling down below. These results are conducted for each city cluster. We have to mention, there are both linear and nonlinear multiple regression models. The resulting regression models are significant, and their residuals are distributed according to the normal distribution. Tests for the presence of multicollinearity showed negative results, which rejects the hypothesis of correlation between the regressors of the model. Thus, table 2 shows result down below.

Table 2.

Multiple regression models

4. Conclusion

This study suggested new look on the city development strategy. This analysis helped to visualize the clusters on the city map, and this study was conducted for the city of Krasnoyarsk for the first time. It is important to mention that the cluster analysis and multiple regression functions were formed on the data base of 9.3 thousand observations, which indicates a high representativeness. Moreover, according to the hedonic price formation theory, external factors were included in the analysis in addition to internal ones. Each multiple regression model has a high square of determination, which makes it possible to predict the price per square meter on the primary real estate market. It should also be noted the binding to the geographical component. In the future, similar studies should be carried out for all regional capitals of the Yenisei Siberia in order to increase the scale of the study.

References:

1. Rosen, S. (1974). Hedonic Prices and Implicit Markets: Product Differentiation in Pure Competition. Journal of Political Economy, 82, 34-55.
2. Arevera. Real-estate agency : Database <arevera.ru>
3. Etagi. Real-estate agency : Database <kras.etagi.com>
4. Unified Database of Russian developers. The biggest database of primary real estate and Russian developers <erzrf.ru>