影响四川省房地产业发展的因素分析 “民以食为天，以住为地”，房地产在国民经济生活中起着重要的作用.房地产业的发展对国民经济发展有着巨大而深刻的影响.因此，审慎分析房地产市场的现状及其基本走势,深入探讨影响房地产业发展的主要因素，对促进房地产业事业持续重要. 一、对房地产市场的基本判断 ２０００～２００３年间，我国房地产投资连续四年增幅在２０％以上，２００３年房地产投资增长率高达２９．７％，投资规模首次突破１万亿元，占全社会固定资产投资的１８．３％占ＧＤＰ的８．７％。 在投资规模不断增长的同时，商品房的销售率也在稳步上涨。然而，在总体看好的情况下，局部地区投资增长过快、供求结构失衡的问题仍十分突出。除区域性问题外，投资结构也不容忽视。近几年，全国普通住宅销售率连续四年在８０％以上，且不断走高，但高档住宅、经济适用房、办公楼宇和商厦的销售率却都比较低，导致办公楼和商厦的大量闲置。简言之，此轮房地产投资热，有合理的因素，如中国经济的高速增长、城市化进程的加快、居民收入水平的提高、消费的升级换代、大量外资的进入和劳动力的流动形成的市场有效需求等，但是，政府人为造市、无视地区市场环境的盲目大干快上，导致奢靡之风、非理性投资泛滥，由此引发的投资热和结构失衡，对房地产业来说则是贻害长远。 二、选定变量进行计量经济学的分析 我们选定四川省房地产开发总值(单位:亿元)作为被解释变量Ｙ,同时一共选定了八个解释变量: Ｘ1—四川省每年税收总额(单位:万元) Ｘ2—四川省每年储蓄存款总额(单位:亿元) Ｘ3—四川省建筑材料工业品出厂价格指数(单位:％) Ｘ4—四川省原材料燃料和动力购进价格指数(建筑材料类)(单位:％) Ｘ5—四川省每人每年可支配收入(单位:元) Ｘ6—四川省ＧＤＰ(单位:亿元) Ｘ7—货币供应量(单位:亿元) Ｘ8—贷款利率(3年-5年期)(单位:％) 年份 Y X1 X2 X3 X4 X5 X6 X7 X8 1900 5.18 256513 698.15 101.2 100 1862.36 890.95 15293.4 11.7% 1991 6.09 323350 813.47 96.4 118.2 2132.56 1016.31 19349.9 11.7% 1992 14.19 486213 904.68 92.4 139.3 2988.21 1177.27 25402.2 11.7% 1993 34.87 658821 988.17 87.2 168.5 3422.17 1486.08 34879.8 11.7% 1994 50.57 869451 1024.21 98.5 215.1 3846.22 2001.41 46923.5 11.7% 1995 79.96 1052013 1148.59 103.7 273.3 4002.91 2504.95 60750.5 11.7% 1996 90.62 1298407 1521.97 140.5 294.4 4406.09 2985.15 76094.9 11.7% 1997 100.26 1525792 1841.22 97.0 354.6 4763.26 3320.11 90995.3 9.9% 1998 120.60 1660286 1984.54 95.3 417.2 5127.08 3580.26 104498.5 7.65% 1999 142.50 1778979 2351.21 94.0 417.2 5477.89 3711.61 119897.9 6.03% 2000 195.97 1946319 2693.17 94.7 444.7 5894.27 4010.25 134610.3 6.03% 2001 268.15 2181466 3123.39 97.3 470.7 6360.47 4421.76 158301.9 6.03% 构造模型: （数据见下页） Ｙ＝0+1Ｘ1+2Ｘ2+3Ｘ3+4Ｘ4+5Ｘ5+6Ｘ6+7Ｘ7+8Ｘ8 对数据进行分析: Dependent Variable: Y Method: Least Squares Date: 12/22/04 Time: 16:40 Sample: 1990 2002 Included observations: 13 Variable Coefficient Std. Error t-Statistic Prob. C 9.506190 61.47340 0.154639 0.8846 X1 -0.000108 0.000146 -0.737120 0.5019 X2 -1.32E-06 1.34E-06 -0.981761 0.3818 X3 0.515279 0.894020 0.576362 0.5952 X4 -1.019800 0.958646 -1.063793 0.3474 X5 -0.019462 0.037130 -0.524162 0.6279 X6 -0.154027 0.062463 -2.465914 0.0692 X7 0.007994 0.001154 6.925113 0.0023 X8 9.830840 3.240301 3.033929 0.0386 R-squared 0.995226 Mean dependent var 115.8538 Adjusted R-squared 0.985677 S.D. dependent var 114.4145 S.E. of regression 13.69309 Akaike info criterion 8.277619 Sum squared resid 750.0024 Schwarz criterion 8.668738 Log likelihood -44.80453 F-statistic 104.2251 Durbin-Watson stat 3.388729 Prob(F-statistic) 0.000226 对数据的多重共线性进行分析: X1 X2 X3 X4 X5 X6 X7 X8 X1 1.000000 0.992449 -0.490698 -0.579667 0.926422 0.942814 0.985242 -0.808622 X2 0.992449 1.000000 -0.482882 -0.568568 0.920679 0.932850 0.975851 -0.819597 X3 -0.490698 -0.482882 1.000000 0.900281 -0.486587 -0.511065 -0.489324 0.424923 X4 -0.579667 -0.568568 0.900281 1.000000 -0.638015 -0.648189 -0.604672 0.455466 X5 0.926422 0.920679 -0.486587 -0.638015 1.000000 0.995972 0.969636 -0.620918 X6 0.942814 0.932850 -0.511065 -0.648189 0.995972 1.000000 0.981303 -0.655603 X7 0.985242 0.975851 -0.489324 -0.604672 0.969636 0.981303 1.000000 -0.749205 X8 -0.808622 -0.819597 0.424923 0.455466 -0.620918 -0.655603 -0.749205 1.000000
可见解释变量间存在多重共线性,对此我们进行修正,采用逐步回归：
Y与X7的拟合效果最好 Dependent Variable: Y Method: Least Squares Date: 12/22/04 Time: 16:54 Sample: 1990 2002 Included observations: 13 Variable Coefficient Std. Error t-Statistic Prob. C -47.59996 17.42865 -2.731133 0.0195 X7 0.001982 0.000178 11.15827 0.0000 R-squared 0.918823 Mean dependent var 115.8538 Adjusted R-squared 0.911444 S.D. dependent var 114.4145 S.E. of regression 34.04795 Akaike info criterion 10.03406 Sum squared resid 12751.89 Schwarz criterion 10.12097 Log likelihood -63.22136 F-statistic 124.5069 Durbin-Watson stat 0.631549 Prob(F-statistic) 0.000000
将其余变量逐一引入的如下几个模型: 引入Ｘ6: Dependent Variable: Y Method: Least Squares Date: 12/22/04 Time: 17:01 Sample: 1990 2002 Included observations: 13 Variable Coefficient Std. Error t-Statistic Prob. C 24.27906 25.19510 0.963642 0.3579 X6 -0.090727 0.027503 -3.298789 0.0080 X7 0.004151 0.000670 6.196114 0.0001 R-squared 0.961126 Mean dependent var 115.8538 Adjusted R-squared 0.953351 S.D. dependent var 114.4145 S.E. of regression 24.71162 Akaike info criterion 9.451598 Sum squared resid 6106.641 Schwarz criterion 9.581971 Log likelihood -58.43539 F-statistic 123.6208 Durbin-Watson stat 0.917644 Prob(F-statistic) 0.000000
Ｘ6通过检验,引入Ｘ８： Dependent Variable: Y Method: Least Squares Date: 12/22/04 Time: 17:05 Sample: 1990 2002 Included observations: 13 Variable Coefficient Std. Error t-Statistic Prob. C -80.26395 34.93053 -2.297816 0.0472 X6 -0.143206 0.024342 -5.882998 0.0002 X7 0.005851 0.000676 8.655646 0.0000 X8 11.17218 3.235669 3.452819 0.0072 R-squared 0.983278 Mean dependent var 115.8538 Adjusted R-squared 0.977703 S.D. dependent var 114.4145 S.E. of regression 17.08441 Akaike info criterion 8.761870 Sum squared resid 2626.894 Schwarz criterion 8.935700 Log likelihood -52.95215 F-statistic 176.3999 Durbin-Watson stat 1.513796 Prob(F-statistic) 0.000000
Ｘ8通过检验,引入Ｘ2: Dependent Variable: Y Method: Least Squares Date: 12/22/04 Time: 17:07 Sample: 1990 2002 Included observations: 13 Variable Coefficient Std. Error t-Statistic Prob. C -50.80815 26.01905 -1.952729 0.0866 X2 -2.24E-06 6.92E-07 -3.237378 0.0119 X6 -0.162752 0.018028 -9.027545 0.0000 X7 0.007402 0.000673 11.00720 0.0000 X8 8.109910 2.448140 3.312683 0.0107 R-squared 0.992761 Mean dependent var 115.8538 Adjusted R-squared 0.989142 S.D. dependent var 114.4145 S.E. of regression 11.92238 Akaike info criterion 8.078435 Sum squared resid 1137.146 Schwarz criterion 8.295723 Log likelihood -47.50983 F-statistic 274.2850 Durbin-Watson stat 2.434411 Prob(F-statistic) 0.000000 Ｘ2通过各项检验,而其他变量未能通过各项检验,以上变量模型拟合度良好: 利用ARCH检验是否存在异方差 ARCH Test: F-statistic 0.444405 Probability 0.730040 Obs*R-squared 1.818050 Probability 0.611015 Test Equation: Dependent Variable: RESID^2 Method: Least Squares Date: 12/22/04 Time: 17:18 Sample(adjusted): 1993 2002 Included observations: 10 after adjusting endpoints Variable Coefficient Std. Error t-Statistic Prob. C 197.6702 116.5979 1.695315 0.1409 RESID^2(-1) -0.389754 0.411792 -0.946481 0.3804 RESID^2(-2) -0.403048 0.444113 -0.907535 0.3991 RESID^2(-3) -0.091219 0.421512 -0.216410 0.8358 R-squared 0.181805 Mean dependent var 103.1534 Adjusted R-squared -0.227292 S.D. dependent var 136.2191 S.E. of regression 150.9079 Akaike info criterion 13.16039 Sum squared resid 136639.2 Schwarz criterion 13.28143 Log likelihood -61.80196 F-statistic 0.444405 Durbin-Watson stat 1.941317 Prob(F-statistic) 0.730040
Obs*R-squared=1.818050<7.81=x^2 0.05(3) 所以不存在异方差
利用Cochrane-Orcutt迭代法检验 Dependent Variable: Y Method: Least Squares Date: 12/22/04 Time: 17:36 Sample(adjusted): 1991 2002 Included observations: 12 after adjusting endpoints Convergence achieved after 10 iterations Variable Coefficient Std. Error t-Statistic Prob. C -46.53201 27.88659 -1.668616 0.1462 X2 -2.54E-06 9.14E-07 -2.775620 0.0322 X6 -0.167869 0.018630 -9.010731 0.0001 X7 0.007674 0.000789 9.719952 0.0001 X8 7.825420 2.701815 2.896357 0.0275 AR(1) -0.307889 0.472444 -0.651695 0.5387 R-squared 0.992656 Mean dependent var 125.0767 Adjusted R-squared 0.986536 S.D. dependent var 114.3435 S.E. of regression 13.26767 Akaike info criterion 8.315390 Sum squared resid 1056.186 Schwarz criterion 8.557844 Log likelihood -43.89234 F-statistic 162.2015 Durbin-Watson stat 2.025073 Prob(F-statistic) 0.000003 Inverted AR Roots -.31 此时 D1<DW=2.025073< 4-D2 所以模型的自相关性得到修正,不存在自相关 最终我们得到的模型为： Ｙ＝-50.80815－2.24E-06Ｘ2－0.162752Ｘ6＋0.007402Ｘ7＋8.109910Ｘ8 三、经济意义分析 从上述分析可知, Ｘ2—四川省每年储蓄存款总额,Ｘ6—四川省ＧＤＰ,Ｘ7—货币供应量,Ｘ8—贷款利率(3年-5年期)是影响四川省房地产开发总值的最主要的因素。 五.2005年四川房地产业状况分析 1、土地拍卖继续使房地产开发中土地成本上涨，利润降低。 2、建材上涨房产运作压力加大 3、配套设施提高成本上升 4、央行加息房价上升 5、土地增值税的严格征管 6、信贷政策制约房产发展
