“The effect of spatiality on house price increases: Geographical distribution of house price increase rate on the basis of NUTS2 regions”
The housing sector, which often acts as a lever for country economies with the added value and employment opportunities it creates, has a special importance in the economy. Considering the supply and demand dynamics, it is seen that the housing market is affected by many macroeconomic variables (İlhan & Gökçe, 2023), while it is known that urban transformation works, migration and environmental factors that have accelerated recently are among the factors affecting the housing market.
Factors such as the structural features of the houses, their location in the city, and their distance to various focal points are the most important factors in the change of house prices. Accessibility and transportation networks, economic centers and job opportunities, natural resources and landscape, infrastructure and services, investment and development opportunities, regulations and policies affected by the location of geographical regions may be among the factors affecting house price growth.
The aim of this study is to reveal the differences in the house price increase rate on the basis of Turkey's 26 levels compared to NUTS2. In this context, the increase in house prices for the last 12.75 years was investigated on a 26-level basis and the house price increase rates were calculated. Inter-regional comparisons were made by descriptive table, Plots, Kernel Density Estimates, ANOVA/T-test, Global-Local Moran I, Local Geary Global statistical analyses, respectively. 26 Regions were divided into 3 categories: coastal-inland, inland-mountainous, coastal-mountainous, and an attempt was made to interpret the reasons for the increase in housing price rates.
Source: Produced by the author with data provided by the Central Bank of the Republic of Turkey
Source: Konut Fiyat Endeksi . (2023, December 30). Türkiye Cumhuriyet Merkez Bankası EVDS Verinin Merkezi: retrieved from https://evds2.tcmb.gov.tr/index.php?/evds/dashboard/310
DESCRIPTIVE TABLE
The central tendency of the data set is 124,343. The data set ranges from 77,534 to 221,402 values. The width of the distribution of the data set is seen as 143,867. Each piece of data deviates from the mean by 34,037 units.
The fact that the mean and median values are close to each other shows that the data set has a structure close to symmetric or normal distribution. In other words, house price increase rates show a normal distribution.
A low coefficient of variation value (27.3%) indicates that the housing price increase rates is relatively homogeneous or the variability is lower than the average value.
Source: In excel environment Produced by the author with data provided by the Central Bank of the Republic of Turkey
PLOTS
Source: In ArcGIS environment produced by the author with data provided by the Central Bank of the Republic of Turkey
The region with the highest increase rate is TR32 (Aydın, Denizli, Muğla) with 197.95%, and the region with the lowest housing price increase rate is TR90 (Artvin, Giresun, Gümüşhane, Ordu, Rize, Trabzon) with 104.36% in coastal cities.
Source: In ArcGIS environment produced by the author with data provided by the Central Bank of the Republic of Turkey
This situation is TR61 (Antalya, Burdur, Isparta) with 221.40% cities, TR63 (Hatay, Kahramanmaraş, Osmaniye) with 77.53% for inland cities.
Source: In ArcGIS environment produced by the author with data provided by the Central Bank of the Republic of Turkey
We encounter TRC1 (Kilis, Adıyaman, Gaziantep) with 129.46% and TRB1 (Bingöl, Elazığ, Malatya, Tunceli) with 80.10% for mountainous cities, respectively.
Source: In ArcGIS environment produced by the author with data provided by the Central Bank of the Republic of Turkey
KERNEL DENSITY ESTIMATES (KDE)
According to KDE analysis, which is a non-parametric method, we can say that the data set is concentrated in 2 poles. It is seen that some of the regions in the data set have a high house price increase rate, while the other group has a relatively lower rate.
We can say that house price increases have increased relatively slowly, especially in the eastern region, in 12.75 years.
Source: In eviews environment produced by the author with data provided by the Central Bank of the Republic of Turkey
ANOVA & T-TEST
The p and f values reached in the ANOVA test performed to investigate whether there is a significant difference between 3 separate groups are 0.029 and 3.422, respectively. Since the determined significance level is 0.05 > p value=0.029, it is concluded that there is a statistically significant difference between the groups. Since the f value, which measures the difference in variance, is significantly larger than 1, it is concluded that the difference is significant between the coastal, inland and mountainous groups.
Among the p values obtained as a result of the t-test, p valuecoastal−inland〖"p value" 〗_"coastal-inland" =0,140 indicates that there is no significant difference between the housing price increase rates of the regions in this category, but since the p-value of the other 2 categories is low, housing It can be said that there is a significant difference between the price increase rates.
LOCAL GEARY (Gİ*)
According to the Local Geary (Gi*) analysis, within the scope of the cluster map, it can be said that the cluster formed by the regions with high housing price increase rates is located in the southwest, and the cluster formed by the regions with low housing increase rates is located in the east and northeast.
According to the Local Geary (Gi*) analysis, within the scope of significance map, it can be said that TR33, TR72, TRA1, TRB1, TRC1, TRB2 regions show significant differences compared to TR32, TR61, TR90, TRA2, TRC2 regions.
Source: In GeoDa environment produced by the author with data provided by the Central Bank of the Republic of Turkey
GLOBAL LOCAL MORAN I
The result of Global Moran-I analysis, which is a statistical method that measures the presence of spatial autocorrelation (local independence or similarity), was recorded as Moran's I: 0.412726. A positive value indicates that there is a spatial concentration and clustering.
The housing price increase rate of the determined area indicates that it is of similar value to neighboring units. In this context, high-high or low-low clustering is observed.
According to the Local Moran-I analysis, there are 4 types of clusters. In this context, it can be said that the housing price increase rates of the eastern regions and their surroundings are low, while the housing price increase rates of the southwestern regions and their surroundings are high.
Considering the significance map, the gray colored regions cannot form a statistically significant difference or a group among themselves; Depending on the significance level of p=0.001, p=0.01, p=0.05, it can be said that there is a significant difference at a very high level, a high level and a limited level, respectively.
Source: In GeoDa environment produced by the author with data provided by the Central Bank of the Republic of Turkey
In this study, which was created to reveal the differences in the housing price increase rate on a 26-level basis in Turkey compared to NUTS2, the housing price increase rates were calculated by investigating the increase in housing prices for the last 12.75 years on a 26-level basis. As a result of the methods applied respectively, it was revealed that the increase in housing prices varies according to geographical locations.
It is seen that there are many differences between regions and spatiality plays an active role in this regard. In particular, it can be said that the main issue here is the southwest and northeast. It should be emphasized that the reason why the regions located in the southwest have very fast housing price increase rates, while the regions in the northeast have relatively slower housing price increase rates is related to geography.
Within the 3 categories determined, it can be said that the mountain is actually a determining feature in the housing price increase rate, and that the environmental and natural values of the regions within the scope of their geographical location and factors such as migration also play a role in the mentioned housing price increase rate.
As a result, it was concluded that the housing price increase rate of 26 regions varies depending on geography and spatiality, in line with various analyses.
