1、房地产影响因素分析房地产影响因素分析（背景）2002年以来，我国商品房销售额大幅攀升 带动了房地产开发和城市基础设施投资的新一轮高速增长。通过产业链的传递，进而又拉动钢材、有色金属、建材、石化等生产资料价格的快速上涨，刺激这些生产资料部门产能投资的成倍扩张，最后导致全社会固定资产投资规模过大、增速过快情况的出现。房价过快上涨在推动投资增长过快的同时，已经成为抑制消费的重要因素。房地产价格本身呈自然上涨趋势，房价中长期趋势总是看涨。随着我国经济发展，居民可支配收入提高，民间资金雄厚，大量资金需要寻找投资渠道，而股票市场等投资渠道目前又处于低迷状态，这是房地产投资需求不断扩大的经济背景。强劲的CP
2、I上涨说明当前的房价上涨并非孤立，是有其宏观经济背景的。宏观调控能否有效防止局部行业过热出现反弹，其中的关键就是要继续加强和完善对房地产业的调控。 （引言）国际上关于房地产有一种普遍的观点：人均收入超过1000美元，房地产市场呈现高速发展阶段。欧美等发达国家基本都经历了这样一个阶段。我们这篇论文，主要探讨房地产影响因素分析，主要从人均收入对房地产长期发展的影响阐述。年份 X1 X2 X3 Y 19902551.7361510.16222704.331919911111.2361700.6233.3786.19351992590.59982026.6253.4994.655519932897.0
3、192577.4294.21291.45619943532.4713496.2367.81408.63919953983.0814282.95429.61590.86319964071.1814838.9467.41806.39919973527.5365160.3481.91997.16119982966.0575425.14792062.56919992818.8055854472.82052.620002674.2646279.98476.62111.61720012830.6886859.6479.92169.71920022906.167702.8475.12250.17720033
4、011.4248472.2479.42359.49920043441.629421.6495.22713.878X1=建材成本（元/平方米 ） X2=居民人均收入（元） X3=物价指数 Y=房地产价格（元/平方米）初定模型：Y=c+a1*x1 +a2*x2 +a3*x3+etDependent Variable: YMethod: Least SquaresDate: 06/05/05 Time: 23:04Sample: 1990 2004Included observations: 15VariableCoefficientStd. Errort-StatisticProb. X32.53
5、75780.5904224.2979080.0013X20.1464950.0209686.9865680.0000X1-0.0180160.035019-0.5144470.6171C33.20929118.27470.2807810.7841R-squared0.983094 Mean dependent var1753.317Adjusted R-squared0.978483 S.D. dependent var600.9536S.E. of regression88.15143 Akaike info criterion12.01917Sum squared resid85477.4
6、2 Schwarz criterion12.20798Log likelihood-86.14376 F-statistic213.2186Durbin-Watson stat1.504263 Prob(F-statistic)0.000000一：多元线性回归 Dependent Variable: YMethod: Least SquaresDate: 06/05/05 Time: 23:05Sample: 1990 2004Included observations: 15VariableCoefficientStd. Errort-StatisticProb. X10.3360100.1
7、510842.2239990.0445C792.0169453.44601.7466620.1043R-squared0.275612 Mean dependent var1753.317Adjusted R-squared0.219889 S.D. dependent var600.9536S.E. of regression530.7855 Akaike info criterion15.51016Sum squared resid3662533. Schwarz criterion15.60457Log likelihood-114.3262 F-statistic4.946171Dur
8、bin-Watson stat0.275870 Prob(F-statistic)0.044490Dependent Variable: YMethod: Least SquaresDate: 06/05/05 Time: 23:09Sample: 1990 2004Included observations: 15VariableCoefficientStd. Errort-StatisticProb. X35.5017790.52507510.478090.0000C-486.8605220.1227-2.2117690.0455R-squared0.894128 Mean depende
9、nt var1753.317Adjusted R-squared0.885984 S.D. dependent var600.9536S.E. of regression202.9191 Akaike info criterion13.58706Sum squared resid535290.2 Schwarz criterion13.68146Log likelihood-99.90293 F-statistic109.7903Durbin-Watson stat0.440527 Prob(F-statistic)0.000000Dependent Variable: YMethod: Le
10、ast SquaresDate: 06/05/05 Time: 23:10Sample: 1990 2004Included observations: 15VariableCoefficientStd. Errort-StatisticProb. X20.2363470.01587914.884170.0000C561.997588.563336.3457130.0000R-squared0.944572 Mean dependent var1753.317Adjusted R-squared0.940308 S.D. dependent var600.9536S.E. of regress
11、ion146.8243 Akaike info criterion12.93992Sum squared resid280245.9 Schwarz criterion13.03432Log likelihood-95.04937 F-statistic221.5384Durbin-Watson stat0.475648 Prob(F-statistic)0.000000Dependent Variable: YMethod: Least SquaresDate: 06/07/05 Time: 21:42Sample: 1990 2004Included observations: 15Var
12、iableCoefficientStd. Errort-StatisticProb. X32.3558330.4583405.1399230.0002X20.1500860.0191577.8347140.0000C37.56794114.29910.3286810.7481R-squared0.982687 Mean dependent var1753.317Adjusted R-squared0.979802 S.D. dependent var600.9536S.E. of regression85.40783 Akaike info criterion11.90961Sum squar
13、ed resid87533.98 Schwarz criterion12.05122Log likelihood-86.32207 F-statistic340.5649Durbin-Watson stat1.408298 Prob(F-statistic)0.000000得到结果发现，的系数小，然后对与回归可决系数小，相关性差，剔出这个因素。因为价格更多取决于供需关系。修正之后为：Yc+a2*x2+a3*x3et二：多重线性分析：三个表如上： X2 与X3 存在多重共线性，1.000000 0.876073 0.876073 1.000000Dependent Variable: YMeth
14、od: Least SquaresDate: 06/05/05 Time: 23:09Sample: 1990 2004Included observations: 15VariableCoefficientStd. Errort-StatisticProb. X35.5017790.52507510.478090.0000C-486.8605220.1227-2.2117690.0455R-squared0.894128 Mean dependent var1753.317Adjusted R-squared0.885984 S.D. dependent var600.9536S.E. of
15、 regression202.9191 Akaike info criterion13.58706Sum squared resid535290.2 Schwarz criterion13.68146Log likelihood-99.90293 F-statistic109.7903Durbin-Watson stat0.440527 Prob(F-statistic)0.000000Sample: 1990 2004Included observations: 15VariableCoefficientStd. Errort-StatisticProb. X20.2363470.01587
16、914.884170.0000C561.997588.563336.3457130.0000R-squared0.944572 Mean dependent var1753.317Adjusted R-squared0.940308 S.D. dependent var600.9536S.E. of regression146.8243 Akaike info criterion12.93992Sum squared resid280245.9 Schwarz criterion13.03432Log likelihood-95.04937 F-statistic221.5384Durbin-
17、Watson stat0.475648 Prob(F-statistic)0.000000由于引入物价指数改善小，所以模型仅一步改进为：Yc+a2*x2et三：异方差检验： ARCH Test:F-statistic1.315031 Probability0.335173Obs*R-squared3.963227 Probability0.265462Test Equation:Dependent Variable: RESID2Method: Least SquaresDate: 06/05/05 Time: 23:46Sample(adjusted): 1993 2004Included
18、observations: 12 after adjusting endpointsVariableCoefficientStd. Errort-StatisticProb. C22737.9410296.612.2082950.0582RESID2(-1)0.2419520.3831440.6314930.5453RESID2(-2)-0.3277690.404787-0.8097340.4415RESID2(-3)-0.2737200.378355-0.7234490.4900R-squared0.330269 Mean dependent var16705.23Adjusted R-sq
19、uared0.079120 S.D. dependent var18205.33S.E. of regression17470.29 Akaike info criterion22.63559Sum squared resid2.44E+09 Schwarz criterion22.79723Log likelihood-131.8136 F-statistic1.315031Durbin-Watson stat1.842435 Prob(F-statistic)0.335173临界值所以无异方差White Heteroskedasticity Test:F-statistic0.159291
20、 Probability0.854522Obs*R-squared0.387928 Probability0.823687Test Equation:Dependent Variable: RESID2Method: Least SquaresDate: 06/05/05 Time: 23:46Sample: 1990 2004Included observations: 15VariableCoefficientStd. Errort-StatisticProb. C31063.2822612.201.3737400.1946X2-5.0557549.640127-0.5244490.609
21、5X220.0004210.0009070.4646050.6505R-squared0.025862 Mean dependent var18683.06Adjusted R-squared-0.136494 S.D. dependent var18673.13S.E. of regression19906.77 Akaike info criterion22.81236Sum squared resid4.76E+09 Schwarz criterion22.95397Log likelihood-168.0927 F-statistic0.159291Durbin-Watson stat
22、1.357657 Prob(F-statistic)0.854522临界值无异方差。四：自相关分析：查表的存在自相关广义差分法修正：Dependent Variable: DYMethod: Least SquaresDate: 06/06/05 Time: 00:18Sample(adjusted): 1991 2004Included observations: 14 after adjusting endpointsVariableCoefficientStd. Errort-StatisticProb. DX20.1820860.0349185.2146550.0002C236.558
23、963.273883.7386500.0028R-squared0.693820 Mean dependent var544.1620Adjusted R-squared0.668305 S.D. dependent var148.7133S.E. of regression85.64840 Akaike info criterion11.86994Sum squared resid88027.77 Schwarz criterion11.96124Log likelihood-81.08959 F-statistic27.19263Durbin-Watson stat1.584278 Pro
24、b(F-statistic)0.000217得出：回归后可决系数降低，考虑其他方法。1迭代法：表：发现可决系数提高，统计量提高，已经无自相关。结论：YbY（）*（）a2*（x2b*x2（）+et由下表的b=0.681C=561.9975 a2=0.236347 179.2772Y*= Y0.681Y（） X*= x20.681*x2（）Y*=179.2272 +0.2363X*+etMethod: Least SquaresDate: 06/07/05 Time: 20:57Sample(adjusted): 1991 2004Included observations: 14 after a
25、djusting endpointsVariableCoefficientStd. Errort-StatisticProb. E20.6805090.1776963.8296240.0024C11.6877324.888250.4696080.6471R-squared0.549989 Mean dependent var15.32764Adjusted R-squared0.512488 S.D. dependent var133.2751S.E. of regression93.05539 Akaike info criterion12.03583Sum squared resid103
26、911.7 Schwarz criterion12.12712Log likelihood-82.25081 F-statistic14.66602Durbin-Watson stat1.313042 Prob(F-statistic)0.0023972改进模型方程（对数法，然后用迭代法）：LL（）（）（LL（）可决系数很高，F统计量相对1中也有提高，DW=1.811.361无自相关。Dependent Variable: LYMethod: Least SquaresDate: 06/06/05 Time: 10:24Sample(adjusted): 1991 2004Included o
27、bservations: 14 after adjusting endpointsConvergence achieved after 7 iterationsVariableCoefficientStd. Errort-StatisticProb. LX20.5862030.1002435.8477990.0001C2.5258100.8823502.8625940.0154AR(1)0.5671440.2204572.5725890.0259R-squared0.980054 Mean dependent var7.460096Adjusted R-squared0.976428 S.D. dependent var0.351331S.E. of regression0.053941 Akaike info criterion-2.814442Sum squared resid0.032006 Schwarz criterion-2.677501Log likelihood22.70109 F-statistic270.24
