In economics, the three elements of production are presented as land, labor, and capital, and production is optimized by an efficient combination of these three elements. Among these factors, land cannot be moved, and labor also takes some time to move, while money as direct capital in modern society can be moved in real time for higher returns. Therefore, assuming that capital for housing investment remains constant, if housing prices in a specific region rise and investment preferences decrease, funds flow into other regions, which are perceived to have a high return on housing investment.
The housing purchasing power index, an indicator of housing consumption capacity of housing consumers, is determined by the level of housing prices in a specific region along with various factors such as financial regulation and income. Recently, the housing market in the metropolitan area has lowered the housing purchasing power of consumers or investors due to soaring prices, while the housing purchasing power in the local housing market, which has relatively stable housing prices, is likely to affect the housing transaction market.
This study aimed to derive the impact of changes in the national housing purchasing power index, which is greatly affected by changes in the metropolitan area's housing prices, on the local housing brokerage market with different characteristics from the metropolitan area.
The analysis method used multiple time-series analysis using the Vector Error Correction Model (VECM) as the main variables of this study, such as the national housing purchasing power index, the number of housing sales transactions in Jeollabuk-do, and other input variables were observed in time series.
In order to examine the characteristics and association of variables, descriptive statistics and variables were first examined, and then cointegration tests were performed to determine the normality of input variables through a unit root test, and to select a model through understanding the long-term equilibrium relationship between variables. In addition, based on the results of the Grandeur causal relationship analysis, a vector error correction model was constructed by setting the variable input order considering the causality between variables. Finally, the impact form for the number of intermediaries was identified through the impact response analysis according to lag, and the explanatory power of each variable was estimated through the variance decomposition analysis of the prediction error. The main analysis results of this study are divided into impact response analysis and predictive error variance decomposition analysis that grasp the degree and shape of the variable by the characteristics of the analysis model.
As a result of the shock response analysis, the shock response to the number of home sales transactions in Jeollabuk-do was lowered from time difference 2, and the shock response to the number of home sales transactions was in the order of gross domestic product, Jeollabuk-do apartment sales price index, Jeollabuk-do population, national housing purchasing power index, total currency, interest rate, and consumer price index.
The predictive error variance solution is similar to the shock response analysis, and the change in the number of housing sales transactions was found to have a somewhat low explanatory power over the entire period, and other variables increased due to the time difference. The variables with high explanatory power were in the order of gross domestic product, Jeollabuk-do apartment sales price index, Jeollabuk-do population, national housing purchasing power index, total currency volume, interest rate, and consumer price index. In particular, the explanatory power of gross domestic product, an indicator of economic growth, continues to increase as the lag increases, while the national housing purchasing power index has a higher explanatory power of initial lag but decreases as the lag increases.