Neighborhood House Price Index,Housing Investment Yield Fluctuation And Micro Housing Market Efficiency

The real estate industry is not only the basic industry of the national economy,but also the pillar industry.It not only plays a leading role in the development of national economy,but also is the necessity of people’s life stability.Since 2015,China’s real estate market has presented "irrational prosperity",which not only makes the housing lose the attribute of public security,but also squeeze the investment in other fields under the housing insurance effect of "low risk and high return".In order to stabilize and healthy development of China’s real estate market,establish a long-term mechanism for the real estate industry and return it to the consumption attribute of providing housing security for people,the central economic work conference at the end of 2016 proposed for the first time that "the house is used for living,not for speculation",and further required "to implement the long-term adjustment of one city,one policy according to the city,and the main responsibility of the city government "Control mechanism" is a new deployment and new requirements made in view of the new situation and problems in the current real estate market.Therefore,for each city,it is a top priority to recognize the effectiveness of its real estate market,formulate targeted policies according to local conditions,and test the policy effect in real time.House price index is a useful tool to help the government to monitor the trend of urban house price.However,the construction method of China’s housing price index is still at a relatively basic stage.The official use of house price is not constant in quality.There is a large estimation deviation in the establishment of urban internal sub market index,which can’t meet the requirements,and some sub markets with small transaction volume can’t even be estimated.The state-space model based on nonparametric Bayesian clustering proposed in this paper can solve the above problems and establish the housing price index with community as the research object.Compared with the existing index construction methods,it has smaller estimation bias and prediction bias.This paper takes the micro sub market of Shanghai’s new housing market and secondhand housing market as the research object,and establishes 479 community price indexes within the city.Moreover,the spatial heterogeneity and spatial discontinuity of the dynamic changes of urban house prices are observed through non-parametric Bayesian clustering.On the basis of the large community housing price index,this paper further explores the efficiency and heterogeneity of micro housing market by applying the theory of market efficiency and housing dual attribute.The conclusion is that the new housing market is superior to the second-hand housing market in terms of weak efficiency test or dual attributes.Through the spatiotemporal validity test,this paper finds that both the new housing market and the second-hand housing market in Shanghai do not satisfy the weak efficiency hypothesis on the macro level,but more than 80% of the communities in Shanghai are weak efficient.Such test results are consistent with the Samuelson efficient market hypothesis,that is,the macro market is always inefficient,while the micro market is always efficient.Therefore,it is necessary to study the micro market.Through the analysis of the efficiency of the micro market,it can accurately reflect the problems and help the policy makers to regulate reasonably.Finally,through the analysis of the dual attributes of the micro housing market,it is found that from 2016 to 2019,the Shanghai housing market is dominated by investment attributes,and the policy of "housing without speculation" has a significant inhibitory effect on the investment motivation of Shanghai housing market,but the policy effect of the second-hand housing market is not satisfactory.This paper can be divided into four parts,a total of six chapters.The first part is the introduction and theory part,focusing on the practical and theoretical background of this study.The second part is composed of the third chapter,which mainly uses the state space model based on Bayesian nonparametric algorithm to construct the community housing price index.The third part is the empirical application,including the fourth and fifth chapters.It mainly studies the effectiveness and dual attributes of Shanghai City Based on the community house price index constructed in front.The fourth part is the sixth chapter,which summarizes the research conclusions,puts forward the corresponding policy recommendations,and points out the limitations of the paper and the direction of further research in the future.The specific chapters are as follows:The first chapter is the introduction,which is divided into five summaries,focusing on the background and significance of the paper,research ideas and research methods,the overall content of the paper,the main contributions and shortcomings of the study.The first chapter introduces the original intention and significance of the research,puts forward the feasible research scheme,and points out the shortcomings.This chapter is the outline of the whole thesis.The second chapter is the theoretical basis and literature review,which summarizes the previous research results on the construction method of housing index and the convenience of housing market investment machine estimation.This chapter has into two parts.The theoretical basis part briefly explains the classical theories involved in this paper,including house price theory,efficient market hypothesis,central geographic theorem and asset pricing model.In the literature review part,three aspects of research literature are reviewed: the existing housing price index construction method,real estate investment return rate and housing market efficiency,and the application of the model in economic field.The literature review in the second chapter is the basis of this study.The third chapter constructs the community housing investment index,which combines the magic of characteristic price with machine-learning algorithm,and estimates the community housing investment index of Shanghai from 2016 to 2019 based on the state-space model.This chapter has three parts: elaborate the detailed steps of index construction,present the results of community housing investment index,and finally verify the superiority and robustness of the index.The success of the index construction in the third chapter is the premise of the subsequent real estate effectiveness analysis.The community index is more effective and more robust than the existing index,which is also the unique research perspective of this paper.This paper also weighted the Shanghai community housing index into Shanghai urban housing index,and compared with the characteristic housing price index and the main urban housing price index published by the National Bureau of statistics to verify the rationality of the index construction.The fourth chapter studies the time efficiency of the housing market.Based on the community level,this paper uses the time series method to test whether the Shanghai real estate market has weak efficiency.It is divided into three parts: data description,research model,and result analysis.In the part of the research model,unit root test,auto-correlation test and k-order auto-regressive test are used to verify whether the housing market at the community level is effective in the time dimension.Finally,combined with transportation and geographic information,the paper analyzes the reasons for the difference of market effectiveness in different regions of the city,and makes empirical test to draw a conclusion.At the same time,this chapter also studies the spatial efficiency of the urban housing market.From a more micro perspective,combined with the big data analysis method non-parametric Bayesian clustering analysis,this chapter studies the spatial correlation of Shanghai’s new housing market.The analysis structure of this chapter is consistent with the fourth chapter,which includes three parts: data description,research model,and result analysis.In addition to the non-parametric Bayesian time series clustering analysis,the VAR model is used to test the spatial weak efficiency of the housing market,and the Granger causality test is used to further test the analysis results of the VAR model.Finally,this paper tests whether the Shanghai housing market satisfies Samuelson’s hypothesis about the efficiency of macro market and micro market.The fifth chapter estimates the dual attributes of micro housing market.Based on the real rate of return of the community housing price index,this paper studies the proportion of investment and consumption attributes in Shanghai new house trading market and Shanghai second-hand housing market,and estimates the policy effect of "housing without speculation" by using time series and difference model.After that,we add the characteristic variables of the community to interact with the policy of "no speculation on housing".We find that the property level of the community and the strength of the developers have different effects on the investment attribute in the new housing market.The sixth chapter is the conclusion of this paper and the discussion of policy implications.On the basis of summarizing the main conclusions and contributions of theoretical analysis and empirical test,this paper provides policy recommendations for the policy regulation under the "house is used for living,not for speculation",and the future research content and further research direction are described.This paper combines the traditional econometric model with machine learning and big data mining technology,and based on previous research,proposes a feasible research framework,which solves the problems of housing index construction and micro real estate efficiency evaluation.As the first attempt to combine measurement and machine learning model in index construction,this paper has the following four innovations:Firstly,in terms of research methods,this paper proposes a research framework which combines traditional econometric model with machine-learning and big data mining technology.This research framework overcomes the obstacles of small sample estimation,and uses clustering method to alleviate the "Curse of dimension" problem in big data analysis.The model has the following three advantages in estimating the index: 1.The correlation between communities is estimated by clustering,and the price index at the missing time point is estimated automatically according to the correlation.2.The root mean square error(RMSE)and mean absolute error(MAE)of the framework are lower than those of the existing econometric models in both intra sample prediction and out of sample prediction.3.It can’t only capture the dynamic price change pattern of the real estate market,but also be used for the construction of other market price indexes.Secondly,from the perspective of research,this paper chooses the micro market such as residential district as the research individual to analyze whether there is heterogeneity in the efficiency of market segmentation.This paper focuses on the research scope of market efficiency at the community level,explores the heterogeneity and spatial structure of urban micro market effectiveness,and improves and complements the deficiencies found in the micro-housing market by existing research.At the same time,combined with some characteristics of residential quarters,such as the strength of real estate developers and the level of property in the community,this paper analyzes whether there is heterogeneity of purchase motivation in different communities.Thirdly,in terms of research output,this paper not only has research conclusions and policy recommendations,but also initiated the community housing price index,which will make up for the gap in the construction of China’s community level housing index.At the same time,the index can provide a basis for further study of the micro housing market,and also provide data support for business forecast and house purchase selection.Finally,in the research point of view,this paper proposes for the first time that the policy of "housing without speculation" can change the investment or consumption attributes of the housing market by changing the risk return relationship of the housing market,and tests this hypothesis through the double difference model.The results of this study could be used for reference to analyze the property of housing market in the future.