Neighborhood Housing Prices: The Precise Distinction of Homogeneous Submarkets based on Chow and Wald Tests

Kyoun-Sup Chung,Wim P.M. Vijverberg 한국주택학회 2003 주택연구 Vol.11 No.1

This paper examines differences in housing prices among neighborhoods by means of hedonic housing price equations. For the Dallas area, the oommonly made assumption of a single housing market across the entire metroplex is soundly rejected, but the precise distinction of homogeneous submarkets or neighborhoods depends on the estimation strategy, in particular whether one properly accounts for heteroskedasticity. While this paper uses the housing price equation to define neighborhoods, there in no suggestion that we believe there is a uniquely defined concept of a neighborhood. There are other social and spatial concepts of neighborhoods with diffrtent boundaries that are far more relevant to other behavioral models and to other policy issues, where a housing neighborhood simply would not be informative and therefore would not apply: one may think of schooling, health care provision, crime, social cohesion, job markets, etc, Whenever such neighborhoods are determined by statistical procedures, induding regression analysis, the insights gained in this paper apply. J.E.L. Classification Number: R10.R21 Keywords:Housing prices, Neighborhoods, Heteroskedasticity

The Comparison of the Efficient Estimators : OLS, WLS, and Spatial Econometric Model

Kyoun-Sup Chung 한국비교정부학회 2011 한국비교정부학보 Vol.15 No.3

The Application of the Efficient Housing Price Estimation with and/or without Heteroskedasticity

Kyoun-Sup Chung(정건섭) 한국주택학회 2009 주택연구 Vol.17 No.2

이분성의 존재유무의 검증도 없이, 모수의 안정성이 확보되지 않은 행정단위의 지역구분 또는 인위적인 결합은 그 모수 추정에 편기 및 비효율을 초래함으로써 여기에서 유추되는 T-검정 및 F-검정도 그 신뢰성을 믿을 수 없다 하겠다. 또한 그 모수 추정에 대한 편기 및 비효율의 방향도 변수들의 상관관계에 따라 모수의 추정값 및 분산값이 과대평가 되었는지 과소평가 되었는지 구분하기가 실제적으로는 매우 어렵다. 즉, 연구지역에서 최소단위지역의 자료가 동질성을 갖고 있지 않다면 여기서 이루어지는 공간적 병합 및 분할은 잘못된 추론을 가져올 수 있다. 따라서 이러한 지역적 이분성 및 공간적 모수의 불안정성을 모형에 적절하게 통제하지 않으면 여기에서 도출된 어떠한 추론 및 예측도 신뢰의 정도가 약하거나 의심스럽게 되는 것이다. 주택하위시장은 다른 하위시장과는 이질적이면서도 동일 하위시장 내에서는 대체적인 주택시장을 의미하고 있다. 따라서 지역시장에 대한 주택정책 수립 시 상대적으로 동질적인 시장을 파악할수 있다면 주택공급에 따라 각 하위시장별 파급효과를 평가할 수 있어서 주택정책의 실효성을 높이는 중요한 정보가 될 수 있을 것이다. 이는 주택하위시장이 존재한다면 그 가격 추정도 그 주택하위시장의 지역특성에 맞는 가격예측이 필요로 되어지며 이를 기반으로 한 지역정책이 수립되어야 한다는 것을 의미한다고도 볼 수 있다. 따라서 지역적 여건과 특성가격의 추정값이 판이하게 다름에도 불구하고 획일적인 주택가격정책은 반드시 지양되어야 할 것이다. 이는 정확한 또는 논리적 방법적으로 근거한 주택가격 추정을 기본 토대로 하여 지역에 맞는 정책이 수립되어야 함을 함축한다고 하겠다. As far as the heteroskedasticity is concerned, it will cause the estimated variance of the coefficient to influence both direction, upward and downward. In other words, the estimated of variance of the coefficient can be overestimated or estimated depends on the nature of causing factors in the model. If we persist in using the usual testing procedures despite heteroskedasticity, whatever conclusions we draw or inferences we make be very misleading. As a result, we can no longer rely on the conventionally computed confidence intervals and the conventionally employed t and F tests. If there is not a homogeneous spatial process underlying the data, any aggregation will tend to be misleading. This aspect of modifiable area unit problem should be considered as a specification issue related to the form of spatial heterogeneity, and not solely as an issue determined by the spatial organization of the data. This paper addresses the question of homogeneity across a housing market. The effect of heteroskedasticity on the test for homogeneity across samples has concentrated on the case of heteroskedasticity between samples but not within samples in the previous research. An assumption of between-sample heteroskedasticity explicitly states that the factor that divides the samples is the only one that generates differences in the spread of the disturbances. Such an assumption may be reasonable in some circumstances after all, the slopes of the equation are presumed to differ as well. The theory usually offers little guidance about the nature of the distribution of the disturbance; however, assuming that heteroskedasticity exists but only as a function of sample-dividing factor is rather specific and too restrictive in many circumstances. Indeed, much of the present-day research on heteroskedasticity is focused on allowing generality and enhancing flexibility of estimation and testing procedures. Typically, a submarket is defined as a set of dwellings that are reasonable close substitutes for one another, but relatively poor substitutes for dwellings in other submarkets. It is reasonable to set the regional housing policy based on the concrete procedure or statistical method such as housing market segmentation; in addition, the relevant information on the homogeneous housing submarkets will be helpful in order to estimate the spread effect on the regional housing submarkets.

The Regional Homogeneity in the Presence of Heteroskedasticity

Chung, Kyoun-Sup,Lee, Sang-Yup Korean System Dynamics Society 2007 한국시스템다이내믹스 연구 Vol.8 No.2

An important assumption of the classical linear regression model is that the disturbances appearing in the population regression function are homoskedastic; that is, they all have the same variance. If we persist in using the usual testing procedures despite heteroskedasticity, what ever conclusions we draw or inferences we make be very misleading. The contribution of this paper will be to the concrete procedure of the proper estimation when the heteroskedasticity does exist in the data, because the quality of dependent variable predictions, i.e., the estimated variance of the dependent variable, can be improved by giving consideration to the issues of regional homogeneity and/or heteroskedasticity across the research area. With respect to estimation, specific attention should be paid to the selection of the appropriate strategy in terms of the auxiliary regression model. The paper shows that by testing for heteroskedasticity, and by using robust methods in the presence of with and without heteroskedasticity, more efficient statistical inferences are provided.

The Regional Homogeneity in the Presence of Heteroskedasticity

( Kyoun Sup Chung ),( Sang Yup Lee ) 한국시스템다이내믹스학회 2007 한국시스템다이내믹스 연구 Vol.8 No.2

An important assumption of the classical linear regression model is that the disturbances appearing in the population regression function are homoskedastic; that is, they all have the same variance. If we persist in using the usual testing procedures despite heteroskedasticity, what ever conclusions we draw or inferences we make be very misleading. The contribution of this paper will be to the concrete procedure of the proper estimation when the heteroskedasticity does exist in the data, because the quality of dependent variable predictions, i.e., the estimated variance of the dependent variable, can be improved by giving consideration to the issues of regional homogeneity and/or heteroskedasticity across the research area. With respect to estimation, specific attention should be paid to the selection of the appropriate strategy in terms of the auxiliary regression model. The paper shows that by testing for heteroskedasticity, and by using robust methods in the presence of with and without heteroskedasticity, more efficient statistical inferences are provided.

부산시 실거래 주택매매 가격을 이용한 공간계량모형의 적합도 비교연구

정건섭 ( Kyoun Sup Chung ),김성우 ( Sung Woo Kim ),이양원 ( Yang Won Lee ) 한국지리정보학회 2012 한국지리정보학회지 Vol.15 No.1

주택시장 분석에 널리 사용되는 헤도닉 방법은 OLS(ordinary least squares) 모형을 이용하는데, 이는 오차가 독립적이며, 평균이 0이고, 분산이 일정하다는 가정에 기초한다. 그러나 공간자기상관이 존재할 경우에는 이러한 가정에 위배되며, 공간효과를 제대로 반영하지 않으면 왜곡된 추정결과를 가져오게 된다. 최근 이에 대한 대안으로 공간계량모형이 도입되고 있는데, 이 연구에서는 OLS 모형과 공간계량모형의 적합도를 비교·평가하고자 한다. 부산시 실거래 주택매매 가격자료를 이용하여 분석한 결과, OLS를 이용한 기존의 헤도닉 모형보다는 공간자기상관을 고려한 공간계량모형들이 보다 설명력이 높았다. Dubin이 제시한 기준과 Log Likelihood 기준을 통해 볼 때 공간계량모형 중에서는 공간자기회귀모형(spatial autoregressive model: SAR)모형의 적합도가 높은 것으로 나타났다. 이를 통해 주택가격에 있어서의 공간효과를 확인할 수 있었으며, 재건축 추진여부가 아파트 매매가격에 매우 큰 영향을 미침을 알 수 있었다. 또한 적절한 공간계량모형의 선택은 정부의 주택정책에 있어서도 매우 중요하다고 하겠다. The OLS(ordinary least squares) method is widely used in hedonic housing models. One of the assumptions of the OLS is an independent and uniform distribution of the disturbance term. This assumption can be violated when the spatial autocorrelation exists, which in turn leads to undesirable estimate results. An alterative to this, spatial econometric models have been introduced in housing price studies. This paper describes the comparisons between OLS and spatial econometric models using housing transaction prices of Busan, Korea. Owing to the approaches reflecting spatial autocorrelation, the spatial econometric models showed some superiority to the traditional OLS in terms of log likelihood and sigma square(σ2). Among the spatial models, the SAR(Spatial Autoregressive Models) seemed more appropriate than the SAC(General Spatial Models) and the SEM(Spatial Errors Models) for Busan housing markets. We can make sure the spatial effects on housing prices, and the reconstruction plans have strong impacts on the transaction prices. Selecting a suitable spatial model will play an important role in the housing policy of the government.

그랜저인과분석을 통한 매매와 전세시장의 주택가격 결정구조 분석

정건섭(Kyoun-Sup Chung),김성우(Sung-Woo Kim),이상엽(Sang-Yup Lee) 한국정책분석평가학회 2011 政策分析評價學會報 Vol.21 No.2

This study is trying to improve the prediction and to find out the determination of housing market price using CSI(Consumer Sentiment Index). In order to do so, the analysis of our study is used the macro economic variables such as manufacture performance, non-manufacture performance, employment rate, mortgage interest rate, M2, etc., and housing market variables such as housing transaction volume, the number of non-selling apartments, housing market price index, rent market price index, CSI for both rent market and housing market, etc. The results of our analysis are as follows. In the housing market model, the mortgage interest rate and the M2 have an impact on the CSI. And the manufacture performance has a Granger causality to the housing sale CSI. In return, the housing sale CSI has a Granger causality to the housing transaction volume, the number of non-selling apartments, and housing price. Accordingly, the financial measures, such as mortgage interest rate and M2 are necessary to the stabilization of Busan housing market; in addition, mid-long term supply housing plan also is needed due to the low adaptive expectation. In the rent market model, the non-manufacture performance has a Granger causality to the rent CSI, but the financial variables such as mortgage interest rate and M2 have no indication of Granger causality. In return, the rent CSI has a Granger causality to the housing transaction volume and the rent price change. There is no significant impact on the number of non-selling apartments, though. Especially, there exists a short term adaptive expectation in the rent market; for example, the rent price change → the change of CSI → the rent price change. Accordingly, the supply of rent housing, the diverse period of rebuild pland, and the management of empty housing are necessary to the stabilization of Busan rent housing market.

공간적탐색기법을 이용한 부산 주택시장 다이나믹스 분석

정건섭(Kyoun-Sup Chung) 한국콘텐츠학회 2012 한국콘텐츠학회논문지 Vol.12 No.2

본 논문의 목적은 공간적탐색기법을 이용한 부산 주택시장 다이나믹스 분석으로써 MATLAB toolbox M-file을 이용하였다. 본 연구에서 사용된 자료는 2006년부터 2009년 2분기까지 공개된 부산지역 아파트 실거래가 64,530개 자료를 기준으로 법정동을 분류하여 각 평균값을 분석에 이용하였다. 주택시장분석에 많이 이용되는 헤도닉가격 모형은 도시주택경제 분야에서 주택시장 다이나믹스를 설명하는데 강력한 분석기법의 하나이다. 그럼에도 불구하고 전통적인 헤도닉가격 모형은 공간적자기상관의 영향력을 반영하지 않는 단점이 있다. 따라서 본 논문에서는 공간자기상관 관계를 반영한 다양한 공간계량모형, 예를 들어, 공간자기회귀모형(SAR), 공간오차모형(SEM), 일반공간모형(SAC) 등을 보통최소자승법을 이용한 전통적 헤도닉가격 모형과 비교하고자 한다. 이를 위해 결정계수(R²), 분산(σ²), 우도함수(Likelihood)의 값 등의 지표들을 이용하였다. 분석결과 공간자기상관을 고려한 공간계량모형이 전통적 헤도닉모형에 비해 높은 설명력을 보여주고 있다. 공간계량모형에서는 공간오차모형(SEM)과 일반공간모형(SAC)이 공간자기회귀모형(SAR) 보다 우수한 설명력을 보이고 있다. The purpose of this paper is to visualize the housing market dynamics with ESDA (Exploratory Spatial Data Analysis) using MATLAB toolbox, in terms of the modeling housing market dynamics in the Busan Metropolitan City. The data are used the real housing price transaction records in Busan from the first quarter of 2006 to the second quarter of 2009. Hedonic house price model, which is not reflecting spatial autocorrelation, has been a powerful tool in understanding housing market dynamics in urban housing economics. This study considers spatial autocorrelation in order to improve the traditional hedonic model which is based on OLS(Ordinary Least Squares) method. The study is, also, investigated the comparison in terms of R², Sigma Square(σ²), Likelihood(LR) among spatial econometrics models such as SAR(Spatial Autoregressive Models), SEM(Spatial Errors Models), and SAC(General Spatial Models). The major finding of the study is that the SAR, SEM, SAC are far better than the traditional OLS model, considering the various indicators. In addition, the SEM and the SAC are superior to the SAR.

The Distinction of Housing Submarket Using the Efficient Estimator

정건섭(Chung,Kyoun-Sup),이상엽(Lee,Sang-Yup) 한국정책분석평가학회 2010 政策分析評價學會報 Vol.20 No.1

주택하위시장은 다른 하위시장과는 이질적이면서도 동일 하위시장 내에서는 대체적인 주택시장의성격을 띄고 있다. 따라서 지역시장에 대한 주택정책 수립 시 상대적으로 동질적인 시장을 파악할 수있다면 주택공급에 따라 각 하위시장별 파급효과를 평가할 수 있어서 주택정책의 실효성을 높이는 중요한 정보가 될 수 있을 것이다. 이것은 주택하위시장이 존재한다면 그 가격 추정도 그 주택하위시장의지역특성에 맞는 가격예측이 필요로 하며 이를 기반으로 한 지역정책이 수립되어야 한다는 것을 의미한다고도 볼 수 있으며, 이는 정확한 또는 논리적 방법으로 근거한 주택가격 추정을 기본 토대로 하여지역에 맞는 정책이 수립되어야 함을 함축한다. 주택하위시장 구분방법에 관한 선행연구들을 보면 목적은 유사하나 그 방법에 있어 많은 차이들을 보이고 있다. 그러나 인위적인 또는 행정구역별로의 통합보다는 공고한 통계적 방법으로서의 하위시장 구분이 더욱더 정확할 것이다. 따라서 본 연구에서는 이분산 점검을 실시를 통하여 그 결과에 따라 다른 검정방법을(Chow 또는 Wald 검정) 적용하였다. 모수의 안정성이 확보되지 않은 행정단위의 지역구분 또는 인위적인 결합은 그 모수 추정에 편기 및 비효율을 초래함으로써 여기에서 유추되는 T-검정 및 F-검정도 그 신뢰성을 믿기 어렵다. 그 모수 추정에 대한 편기 및 비효율의 방향도 변수들의 상관관계에 따라 모수의 추정값 및 분산값이 과대평가 되었는지과소평가 되었는지 구분하기가 매우 어렵다. 즉, 연구지역에서 최소단위지역의 자료가 동질성을 갖고있지 않다면 여기서 이루어지는 공간적 병합 및 분할은 잘못된 추론을 가져올 수 있다. 따라서 이러한지역적 이분성 및 공간적 모수의 불안정성을 모형에 적절하게 통제하지 않으면 여기에서 도출된 어떠한 추론 및 예측도 신뢰의 정도가 약하거나 의심스럽게 되는데, 본 연구는 이러한 한계를 극복하였다.

The Concrete Procedure of Housing Submarket Distinction with and without Heteroskedasticity

정건섭(Chung Kyoun-Sup),이상엽(Lee Sang-Yup) 한국정책분석평가학회 2004 政策分析評價學會報 Vol.14 No.1

This paper addresses the question of homogeneity across a housing market. Studies often assume that the market is homogeneous within a given geographic boundary such as a metropolitan area or a city. Since such spatial units are drawn up for historical and political reasons, one may question whether it is fair to assume that there is a single competitive market for housing. If there is segmentation. assuming homogeneity leads to uninformative estimates of housing price equations and public policy that lacks proper foundation. The most important conclusions of this paper in terms of methodological point of view can be summarized as follows: (1) Hedonic models of housing prices must be corrected for heteroskedasticity to ensure greater efficiency in the estimation of hedonic prices wherever the heteroskedasticity does exist: (2) There are significant submarket differences in hedonic prices of housing attributes, implying that great care should be taken in the specification of the geographic units for which hedonic models are estimated: and (3) After careful submarket specification, a combination OLS (Ordinary Least Squares) with WLS (Weighted Least Squares) estimates provides the greatest predictive power with respect to estimate of housing prices in the study area. In other words, if there has no evidence of heteroskedasticity. the OLS should be utilized; otherwise. the WLS must be utilized. From the point of view of urban theory, the most important conclusion is that markets are segmented: hedonic prices differ among market segments, reflecting differences in the bid and offer curves for housing. It is recommended the dimensions of this segmentation should be explored further.