A Study on the Influence Factors of Real Estate Prices Based on Econometric Model: A Case of Wuhan Yi-hong XU 1,a, Ke-xin XU 2,* and You-quan CHEN 3

Transcription

1 2016 Inernaional Academic Conference on Human Sociey and Culure (HSC 2016) ISBN: A Sudy on he Influence Facors of Real Esae Prices Based on Economeric Model: A Case of Wuhan Yi-hong XU 1,a, Ke-xin XU 2,* and You-quan CHEN 3 1 Jianghan Universiy, Wuhan, , China 2 Deparmen of mahemaics and saisics he Universiy of Melbourne, VIC, 3010, Ausralia 3 Jianghan Universiy, Wuhan, , China. a xuyihong2208@126.com *Corresponding auhor Keywords: Economeric model, Real esae prices, Influence facors, Error correcion model. Absrac. Based on Economeric Model, his paper analyzes he facors influencing real esae prices, aiming a grasping he mechanism of real esae price changing and hen providing a reference for governmen policy. According o he saisical daa of he real esae in Wuhan from 2000 o 2012, he error correcion model of he real esae price is esablished o analyze he facors. The resul shows ha he facors influencing real esae prices are diversified and complex, and here exiss a long-erm equilibrium relaionship among he loan ineres raes, he populaion and he price of he commercial buildings. Therefore, in he long run, he price of he commercial building will be suppored by he supply and he demand, and he effec of ineres rae is relaively more noable. Inroducion Due o he characers such as large marke demand, long indusrial chain and inegraed wih knowledge-inensive, echnology-inensive and labor-inensiy, he real esae indusry has been he basic indusry promoing he naional economy o grow. Real esae is he leading indusry in naional economy. I has long indusry chain and high relevancy wih oher indusries, which deermines ha he sensiiviy coefficien and influence coefficien of real esae marke are above he average of various indusrial secors in naional economy. However, he real esae marke is no developed in China and he indusry srucure is weak, which caused he real esae prices have been consanly climbing in recen decades and a lo of problems and risks exis in he real esae marke. The facors influencing real esae prices are diversified and complex. We need o analyze from he angle of supply, demand and supply-demand, and esablish error correcion model of real esae prices facors, wih he aim of discussing which facors influence real esae prices, how hey influence, and wheher he sensiiviy coefficien is high on he basis of empirical research model. We hope o undersand he mechanism of real esae prices changing hrough hese analyses. According o he real esae empirical analysis from 2000 o 2012 in Wuhan, he paper esablishes error correcion model of he real esae prices facors in Wuhan, discusses saisical model of long-erm rend and shor-erm flucuaion, and hen proposes some counermeasures o reasonably conrol house prices in Wuhan.

2 Reviews of Domesic and Foreign Research Saus In erms of he facors affecing real esae prices, domesic and foreign scholars have done quie a few researches around relaions beween economic fundamenals facors: Case and Shiller (1990)performed a ime series analysis by using cross-secion regression on quarerly daa of four American meropolis, which proved ha housing coss were posiively correlaed wih percenage of prices, populaion, personal income and excessive profis or price changes in real esae marke several years laer. I also explained single-family-home marke was no effecive. Poerba chose four influencing facors, median housing price, building cos, populaion and income o analyze and drew a conclusion ha he rising urban housing price could be explicaed by changes in rue income and building cos. Clapp and Gaccoo performed a simple OLS regression analysis by using repea sale price index, assessmen of residenial value price index, expeced inflaion rae, unexpeced inflaion raes, employmen change, which proved economic change (populaion, employmen, economy) is capable o predic changes in housing price. This doesn accord wih efficien markes hypohesis.case and Mayer (1996)performed an empirical analysis by using Two Sage Leas Square Mehod on annual daa in Boson, and found he median household income had lile effec on housing price. Malpezzi (1999)expounded housing price was no random and could be parly prediced successfully. Quigley (1999) emphasized ha relevan indexes of economic fundamenal could explain he rend of real esae price; hough in he shor erm hey couldn' explain oo much price flucuaion of he real esae. A home, Zeng Xiang added dummy variable D o represen poliical and economic facors (Midwes o, Easern 1), and described he land acquisiion cos (X2), per capia disposable income (X3), affordable housing sale area (X4) are posiively correlaed wih house price changes, hen he obained a simple regression model as is shown in equaion 1: Y = X X X D (1) Zhou Xiaoyan analyzed he reasons of housing price rise from he supply-demand angel, and hen she pu forward he suggesion ha we can promoe he consrucion and consumpion of low-cos housing hrough propery axes.hu Xiufu (2002) proved ha such facors like sopping welfare housing and land ransacion prices affeced real esae price in China;Wang Shuyun (2005) explained ha i was bank s exreme expansion in he real esae marke ha caused real esae price o rise; Yan Jinhai (2006) indicaed ha land price flucuaion caused housing price flucuaion in he shor-erm, and land price and real esae price influenced reciprocally in he long-erm rend, Zhang Mu (2006) emphasized ha he inflow of foreign fixed invesmen coninued o promoe he rise of real esae price in China; Cao Yulong and Xu Fei indicaed ha he main housing price facors in Shanghai are land ransacion price, compleed consrucion area ha very year, providen fund loan ineres rae of personal housing and loan ineres rae of individual housing; Qi Qian (2011)proved ha house price was mainly influenced by facors such as ineres rae, per capia income, price of land and real esae developmen invesmen hrough empirical research. Qiao Lin and Chu Chunli (2012)analyzed 33 ciies in our counry hrough facor of conribuion rae mehod and drew a conclusion ha populaion and income growh are he mos imporan facors affecing housing price. In addiion, invesmen behavior exising in a cerain exen was also a reason causing he growh of housing price.

3 Empirical Analysis of Influence Facors of Real Esae Prices Real esae price including price of land and cos of building is a diversified economic caegory. The value of building, naural land resource, labor inpu in land combine he value of real esae. Real esae price is affeced by many facors such as economy, poliics, sociey, poliics, regulaions, inernal environmen and surroundings. Error Correcion Model Error correcion model was firs pu forward by Sarger. In 1978, Davison, Hendry, Srba and Yeo pu forward he basic framework of error correcion model, so ECM can also be called DHSY model. ECM is a conrol measure o he long-erm equilibrium relaionship deviaion of shor-erm flucuaion beween variables of coinegraion. The Concep of Coinegraion. The concep of coinegraion was firs pu forward by Engle and Granger in 1987, and i made he regression of nonzero single ineger become meaningful. There is long-erm dependency relaion beween some variables in economic sysem, which is called equilibrium relaionship in economics and is he basis of economeric modeling. The equilibrium relaionship shows ha mechanism forming in economic sysem is sable. When season facors or random disurbance make i deparure from is balance poin, he balance mechanism will adjus o make i reurn o equilibrium sae in he nex insallmen. Bu if he deviaion is enduring, he balance mechanism beween variables will be unsable and he balance relaionship will be desroyed. Coinegraion is he saisic expression of his balance relaionship. Time series { X },{ Y } is wo course of I ( 1), if here exiss β making Y β X become I ( 0), we call { X } and { } Y are coinegraion. Acually, coinegraion means ha some kind of linear combinaion of muliple non-saionary ime series is saionary. Generally, if sequences Y1, Y2,, Yk are all d inegraion, hen here exiss nonzero vecor α = ( α1, α2,, αk ), which make αy I ( d b), when b > 0, Y = ( Y1, Y2,, Y k ), we call Y1, Y2,, Yk is (, ) inegraed vecor. d b coinegraion, denoed as Y CI ( d, b). α is co Error Correcion Model (ECM). The co-inegraion relaionship beween variables shows ha here exiss long-erm and sable equilibrium relaionship, which forms under consan flucuaions of he shor-erm dynamic process. The long-erm equilibrium relaionship beween variables is due o an adjusmen mechanism-error correcion mechanism which deviaion of long-erm relaionship is conrolled wihin limis. There exiss error correcion mechanism in every se of coinegraion sequence which reflecs shor-erm regulaion behavior. For sequence Y, X, error correcion model is shown in formula Y = β + β X λecm + ε (2)

4 ecm represens error correcion erm. In general case, 0 < λ < 1. For error correcion model, Engle and Granger proposed he famous Granger represenaion heorem in If variable X and Y are co-inegraed, he shor-erm non-equilibrium relaionship beween hem can always be formulaed by an error correcion model. To esablish an error correcion model, we need do co-inegraion es on variables. When he variables have long-erm equilibrium relaionship, hey can be error correcion erms, and hen we can esablish shor-erm model by regarding error correcion erm as an explanaory variable ogeher wih oher explanaory variables reflecing shor-erm flucuaion. Empirical Analysis of Influence Facors of Real Esae Prices in Wuhan In he real esae marke, due o he complexiy and diversiy of facors affecing he marke price, we can fully selec all facors o do empirical analysis on real esae prices in Wuhan, so we jus selec possible facors affecing economic fundamenal. Therefore, we selec real esae daa of Wuhan in 2000 o 2012, a oal of 13 years, from hree angles, supply, demand, and supply-demand comparison o analyze and sudy. The second and following paragraph under he firs and second level iles (he Inroducion par is excluded) need o indened for one characer. Please do no forge. Supply Model of Real Esae Prices Facors in Wuhan. From he microscopic poin of view, he more he real esae developmen invesmen is, he larger he invesmen of governmen or developers is, and he greaer he size of he building is. Taking no accoun of he housing vacancy rae, he real esae developmen invesmen is posiively correlaed wih he average price of real esae. Land price index is based on Wuhan base land prices in 2000 as a base period. As one of he focus of all secors of sociey, wheher he land price rise really promoes he housing prices has been a roubling problem. According o daa from Minisry of Land, China's land ransfer amoun in 2009 increased by 63.40% compared o Though on he curren siuaion, he causal relaionship beween growh of land and surge of housing prices don' reach a consensus, bu in he course of China's real esae price flucuaions in recen years, he rise of land prices has a very imporan impac on he flucuaion of real esae prices. I is almos he firs choice for conemporary ciizen o purchase residence by morgage loans. So real esae loan ineres rae is bound o be a facor affecing housing price flucuaions. Compleed residenial area represens he oal area of houses ha has been compleed and can be sold o ciizen. The greaer he area of compleed residenial are, he more he house is, and hen he housing prices may fall. Therefore, we inroduce he following four variables o describe he supply model of real esae prices facors in Wuhan: Real esae developmen invesmen (X1): uni/ yuan; Land price index (X2); RMB loan ineres raes rae (X3); Compleed residenial area (X4): uni / en housand square meers; Y = *X *X + (3) 3 4 ε Then we build a regression model for X5, X6 and X7 o verify he facors ha he demand facors have on he real esae prices in Wuhan. However, we find ha variable X3 is highly relaed o X4, wihou passing he significance esing, so we use Sepwise Regression o obain he supply model of real esae prices facors in Wuhan, as shown in formula 3 and able1 of he regression resul.

5 Table 1. Regression Model of Supply of he Real Esae Prices Facors in Wuhan. he regression resul Prob. F-saisic Prob Variable Sd. Coefficien -Saisic (F-sais Error ic) C X X Demand Model of he Real Esae Prices Facors in Wuhan. From he microcosmic poin of view, he greaer he populaion is, he greaer he demand for housing. The populaion growh is posiively correlaed wih average price of commercial housing. GDP is an objecive indicaor reflecing ciizen s paymen capaciy. Generally speaking, he lager he variable is, he sronger he residens' purchasing abiliy.gdp had grown eighfold from 2000 o 2012, which was he main reason ha simulaed housing consumpion o increase rapidly. GDP is posiively correlaed wih housing price. Similarly, he higher he average disposable income is, he sronger he purchasing abiliy is. The average disposable income is posiively correlaed wih housing price. Therefore, we inroduce hree variable indicaors o describe model of demand of real esae prices facors in Wuhan: Populaion(X5): uni / en housand people GDP(X6): uni / one hundred million yuan Average disposable income(x7): uni /one yuan Then we build a regression model for X5, X6 and X7 o verify he facors ha he demand facors have on he real esae prices facors in Wuhan. However, we find x7 is insignifican and isn' esed by, so we use Sepwise Regression o obain he supply model of real esae prices facors in Wuhan, as shown in formula 4 and able2 of he regression resul. Y = *X *X + (4) 5 6 ε Table 2. The Regression Resul of Demand Model of Real Esae Prices Facors in Wuha. he regression resul Prob. F-saisic Prob Variable Sd. Coefficien -Saisic (F-sais Error ic) C X X Supply-demand Model of Real Esae Prices Facors in Wuhan. (1) Co-inegraion Tes We have buil wo fiing models-supply model and demand model on he real esae prices facors in Wuhan. However, because we need o sudy he real esae prices facors under supply and demand and as he main facors hey have long-erm equilibrium relaionship. So we did uni roo es on all variable, afer confirming hey are he same order inegraion, we did co-inegraion es.

6 Table 3. The Tes Resul of he Real Esae Prices Facors in Wuhan(1). variable criical value The Tes Resul p-value real esae developmen invesmen no land price index no RMB loan ineres raes no Compleed area of residenial no populaion no GDP no average disposable income no average price of commercial buildings no Table 4. The Tes Resul of he Real Esae Prices Facors in Wuhan(2). variable criical value The Tes Resul Whehersai p-value onary or no real esae developmen invesmen Yes land price index Yes I(2) RMB loan ineres raes Yes I(2) Compleed area of I(2) Yes residenial populaion Yes I(2) Whehersaion ary or no original sequence inegraion order I(2) GDP Yes I(2), d>2 average disposable Yes income I(2) average price of Yes commercial buildings I(2) According o able3,4, we can know he original series are all non-saionary series excep GDP, and oher variables are all inegraed of order 2.Therefore, we ge rid of variable X6(GDP) and do co-inegraion es on oher variables. Y = *X *X + (5) Firsly, we adop OLS leas square esimaion on hese variables, and hen use sepwise regression o obain model formula like formula 5 and oupu resul able ε Table 5. Tes Resul of Sepwise Regression. Tes Resul Variable Sd. Prob. F-saisic Prob Coefficien -Saisic Error (F-saisic) C X X Then we es wheher he residual sequence e is saionary or no. The value of ADF is and is less han criical value ,which can show ha variable indexes reserved by sepwise regression, included RMB loan ineres raes, populaion has long-erm equilibrium relaionship(co-inegraion relaionship) wih price of commercial buildings.

7 As we can see from long-erm equilibrium model (formula 5), on he demand side, populaion has long-erm equilibrium relaionship wih price of commercial building. When he populaion adds one uni, he prices of commercial buildings add one uni. The larger populaion is, he more he housing in need is, hereupon, he higher he price of commercial building is. On he demand side, RMB loan ineres rae has long-erm equilibrium relaionship wih price of commercial building. When he RMB loan ineres rae adds one uni, he price of commercial building decreases abou 0.06 uni. The higher he RMB loan ineres rae, he larger he house-purchase burden for residens, he more insufficien he supply is. Therefore in he long erm, price of commercial building in Wuhan is suppored by supply and demand, and he effec of ineres rae is relaively more noable. In order o curb he growh in house price, we can add supply properly o make he price of commercial building fall a lile auomaically under he ineracion of supply and demand. (2) Build ECM In order o reduce mulicollineariy caused by disribued-lag model, we coninue o build ECM o reflec shor-erm regulaion behavior and long-erm equilibrium relaionship beween variables. According EG mehod of Engle-Granger, we can obain he ECM formula-formula 6, and oupu resul able6. Table 6. Tes Resul of ECM. Variable Tes Resul Prob. F-saisic Prob Sd. Coefficien -Saisic (F-saisic Error ) C X X ECM(-1) According o shor-erm regulaion relaion, we can know ECM represens shor-erm deviaion of populaion and RMB loan ineres rae, and is absolue value of coefficien represens he speed ha he price deparure. When shor-erm populaion changes one uni, he price of commercial building in Wuhan will adjus abou 0.13 uni; when he shor-erm RMB loan ineres rae changes one uni, he price of commercial building in Wuhan will adjus abou 0.3 uni; when he previous price of commercial building deparures one uni, he curren price will flucuae abou 0.61 uni. References [1] Li Zhanfeng. Economerics. Beijing: China Saisics Press, 2011: In Chinese. [2] Gao Tiemei. Economeric analysis mehod and modeling reviews applicaion and example Beijing: Tsinghua Universiy press, 2009: In Chinese. (6)

8 [3] Xiang Shujian, Zhang Xueyi. Saisics. Beijing: China Saisics Press, 2011: In Chinese. [4] Wang Bing. An Analysis on Influence Facors of Real Esae Price. Maser Degree Thesis of Beijing Jiaoong Universiy, 2007: In Chinese. [5] Song Chengsun,Zhou Huiping. A Sudy on he Influence Facors of Commercial Housing Prices in Xining Ciy Based on Regression Model. Economic Research Guide, : In Chinese. [6] Sean Holly e al. House Price Since he 1940s: Coinergraion Demography and Asymmeries. Economic Modeling, : [7] Karl E Case e al. Housing Price Dynamics wihin a Meropolian Area [J].Regional Science and Urban Economics, : [8] Cheng Songlin, A predicion and analysis of Commercial Housing Prices in Wuhan based on Grey Theory, Maser Degree Thesis of Cenral China Normal Universiy, 2008:1. In Chinese. [9] Dong Yulong, Hang Zhenfang, Tian Yinli, An Saisical Analysis on he Influence Facors of Commercial Housing Prices in Zhangjiakou. Journal of Hebei Norh College, : In Chinese. [10] Zhou Xiaoyan. Influence Facors and Regulaory Measure of Commercial Housing Prices in Our Counry. Value Engineering, : In Chinese. [11] Zhang Xin, Xu Lei. Periodic Flucuaion of he Real Esae Prices in Our Counry and empirical sudy of Influence Facors. saisical research. In Chinese. [12] Zheng Yongzhi, Non-equilibrium Analysis on demand and supply of Real Esae in Hebei. Maser Degree Thesis of Hebei Universiy, 2009: In Chinese. [13] Zheng Yuanyuan, Song Ge, Li Xuanxiu. An Analysis on he Influence Facors of Commercial Housing Price in Harbin. Journal of Norheas Agriculural universiy, : In Chinese. [14] Zeng Xiang, The Applicaion of Economeric Model in Commercial Housing Price in Our Counry, economic research, : In Chinese. [15] Tu Yifeng. An Analysis on Influence Facors of Housing Price in Chongqing. Maser Degree Thesis of Xiamen Universiy, 2008; 50. In Chinese.