影响我国旅游收入因素的实证分析_第1页
影响我国旅游收入因素的实证分析_第2页
影响我国旅游收入因素的实证分析_第3页
影响我国旅游收入因素的实证分析_第4页
影响我国旅游收入因素的实证分析_第5页
已阅读5页,还剩11页未读 继续免费阅读

下载本文档

版权说明:本文档由用户提供并上传,收益归属内容提供方,若内容存在侵权,请进行举报或认领

文档简介

1、影响我国旅游收入因素的实证分析一经济理论背景摘要:影响一国旅游收入的因素有很多,本文针对我国旅游收入影响因素建立了计量经济模型,并利用Eviews软件对收集到的数据进行相关回归以及多重共线性分析,建立了旅游收入影响因素的模型,分析了影响旅游收入主要因素及其影响程度,并提出了相关政策建议。关键字:旅游收入 旅游收入影响因素二指标选取(附数据表)本研究报告的数据来源于“中国统计年鉴”采集数据的区间为1998年2010年年份旅游收入(亿美元)国内旅游人数(亿人次)入境旅游人数(万人次)居民人均旅游花费(元/天)城镇居民家庭人均收入(元)农牧民家庭人均收入(元)1998120.746.446347.8

2、4345435319821999140.997.197279.563944770.520032000162.247.448344.39426.65129.120382001177.927.848901.29449.55535.919732002203.858.789790.83441.8605120862003174.068.79166.21395.77012.922682004257.3911.0210903.82427.58123.126062005292.9612.1212029.23436.19136.829892006339.4913.9412494.21446.9103583342

3、2007419.1916.113187.33482.61237839532008408.4317.1213002.745111443346562009396.7519.0212647.59535.415849.249382010458.1421.0313376.22598.219109.45919三 实验过程1.模型建立以国家旅游收入为被解释变量,国内旅游人数、入境旅游人数、居民人均旅游花费、城镇居民家庭人均收入、农牧民家庭人均收入作为解释变量建立线性回归模型:Yt=0+1X1t+2X2t +3X3t+4X4t+5 X5t+ ui其中,Yt 旅游收入 X1t 国内旅游人数 X2t入境旅游人数

4、X3t居民人均旅游花费X4t城镇居民家庭人均收入 X5t-农牧民家庭人均收入 0、1、2、3、4、5表示待定系数ui 表示随机误差项2、OLS回归模型Dependent Variable: YMethod: Least SquaresDate: 06/22/12 Time: 20:11Sample: 1998 2010Included observations: 13CoefficientStd. Errort-StatisticProb.C-226.023872.69947-3.1090160.0171X116.7330311.271381.4845580.1812X20.0275790.0

5、073183.7687440.0070X30.0212340.1554050.1366350.8952X4-0.0342030.012779-2.6764670.0317X50.1014000.0439702.3060890.0545R-squared0.992782Mean dependent var273.2423Adjusted R-squared0.987626S.D. dependent var119.3226S.E. of regression13.27328Akaike info criterion8.313420Sum squared resid1233.259Schwarz

6、criterion8.574166Log likelihood-48.03723Hannan-Quinn criter.8.259825F-statistic192.5546Durbin-Watson stat2.296421Prob(F-statistic)0.000000Yt=-226.0238+16.73303X1+0.027579X2+0.021234X3-0.034203X4+0.101400 X5t=(-3.109016)(1.484558)(3.768744)(0.136635)(-2.676467)(2.306089)R2=0.992782 R2=0.987626 F=192.

7、5546 DW=2.2964213 多重共线性及其修正从回归结果的系数以及t值我们可以看出模型可能存在多重共线性,下面我们计算出解释变量的相关系数。解释变量的相关系数矩阵如下: X1X2X3X4X5X110.911740.914090.13820.98568X20.9117410.8180.04020.7719X30.914090.81810.1480.9934X40.13820.04020.14810.3488X50.985680.77190.99340.34881由各相关系数值可知, 解释变量之间都高度相关,模型存在严重的多重共线性。多重共线性的修正采用逐步回归法,来检验并解决多重共线性问

8、题。分别作y对x1、x2、x3、x4、x5的一元回归Dependent Variable: YMethod: Least SquaresDate: 06/22/12 Time: 20:42Sample: 1998 2010Included observations: 13CoefficientStd. Errort-StatisticProb.C-12.1732219.46763-0.6253050.5445X123.672341.50273415.752840.0000R-squared0.957554Mean dependent var273.2423Adjusted R-squared0

9、.953695S.D. dependent var119.3226S.E. of regression25.67651Akaike info criterion9.469668Sum squared resid7252.113Schwarz criterion9.556583Log likelihood-59.55284Hannan-Quinn criter.9.451803F-statistic248.1520Durbin-Watson stat1.022698Prob(F-statistic)0.000000Dependent Variable: YMethod: Least Square

10、sDate: 06/22/12 Time: 20:44Sample: 1998 2010Included observations: 13CoefficientStd. Errort-StatisticProb.C-228.575945.38255-5.0366470.0004X20.0474550.00419211.320130.0000R-squared0.920946Mean dependent var273.2423Adjusted R-squared0.913759S.D. dependent var119.3226S.E. of regression35.04118Akaike i

11、nfo criterion10.09156Sum squared resid13506.73Schwarz criterion10.17848Log likelihood-63.59516Hannan-Quinn criter.10.07370F-statistic128.1454Durbin-Watson stat0.419824Prob(F-statistic)0.000000Dependent Variable: YMethod: Least SquaresDate: 06/22/12 Time: 20:44Sample: 1998 2010Included observations:

12、13CoefficientStd. Errort-StatisticProb.C-443.8537120.1817-3.6931870.0035X31.5826440.2626806.0249970.0001R-squared0.767445Mean dependent var273.2423Adjusted R-squared0.746303S.D. dependent var119.3226S.E. of regression60.10076Akaike info criterion11.17056Sum squared resid39733.11Schwarz criterion11.2

13、5748Log likelihood-70.60864Hannan-Quinn criter.11.15270F-statistic36.30058Durbin-Watson stat0.452934Prob(F-statistic)0.000086Dependent Variable: YMethod: Least SquaresDate: 06/22/12 Time: 20:45Sample: 1998 2010Included observations: 13CoefficientStd. Errort-StatisticProb.C47.3349523.651852.0013220.0

14、706X40.0240250.00226410.612820.0000R-squared0.911026Mean dependent var273.2423Adjusted R-squared0.902938S.D. dependent var119.3226S.E. of regression37.17473Akaike info criterion10.20977Sum squared resid15201.56Schwarz criterion10.29669Log likelihood-64.36353Hannan-Quinn criter.10.19191F-statistic112

15、.6320Durbin-Watson stat0.912764Prob(F-statistic)0.000000Dependent Variable: YMethod: Least SquaresDate: 06/22/12 Time: 20:45Sample: 1998 2010Included observations: 13CoefficientStd. Errort-StatisticProb.C9.09561330.710920.2961690.7726X50.0842610.0090689.2918420.0000R-squared0.886992Mean dependent va

16、r273.2423Adjusted R-squared0.876719S.D. dependent var119.3226S.E. of regression41.89587Akaike info criterion10.44889Sum squared resid19307.91Schwarz criterion10.53581Log likelihood-65.91778Hannan-Quinn criter.10.43102F-statistic86.33833Durbin-Watson stat0.732076Prob(F-statistic)0.000002比较可知X1的修正可决系数

17、最大,应该以X1为基础,顺次加入其他变量逐步回归Dependent Variable: YMethod: Least SquaresDate: 06/22/12 Time: 20:52Sample: 1998 2010Included observations: 13CoefficientStd. Errort-StatisticProb.C-114.934727.52498-4.1756520.0019X114.852112.3167816.4106660.0001X20.0197740.0047364.1755370.0019R-squared0.984529Mean dependent

18、var273.2423Adjusted R-squared0.981434S.D. dependent var119.3226S.E. of regression16.25844Akaike info criterion8.614276Sum squared resid2643.370Schwarz criterion8.744649Log likelihood-52.99279Hannan-Quinn criter.8.587478F-statistic318.1755Durbin-Watson stat1.415696Prob(F-statistic)0.000000Dependent V

19、ariable: YMethod: Least SquaresDate: 06/22/12 Time: 20:53Sample: 1998 2010Included observations: 13CoefficientStd. Errort-StatisticProb.C39.7493588.298660.4501690.6622X125.743763.7636976.8400200.0000X3-0.1697140.281070-0.6038150.5594R-squared0.959047Mean dependent var273.2423Adjusted R-squared0.9508

20、56S.D. dependent var119.3226S.E. of regression26.45186Akaike info criterion9.587704Sum squared resid6997.007Schwarz criterion9.718077Log likelihood-59.32008Hannan-Quinn criter.9.560906F-statistic117.0911Durbin-Watson stat1.188241Prob(F-statistic)0.000000Dependent Variable: YMethod: Least SquaresDate

21、: 06/22/12 Time: 20:53Sample: 1998 2010Included observations: 13CoefficientStd. Errort-StatisticProb.C-73.8196225.68246-2.8743200.0165X151.676439.5625805.4040260.0003X4-0.0293520.009950-2.9500110.0145R-squared0.977305Mean dependent var273.2423Adjusted R-squared0.972766S.D. dependent var119.3226S.E.

22、of regression19.69163Akaike info criterion8.997439Sum squared resid3877.603Schwarz criterion9.127812Log likelihood-55.48335Hannan-Quinn criter.8.970641F-statistic215.3094Durbin-Watson stat2.317272Prob(F-statistic)0.000000Dependent Variable: YMethod: Least SquaresDate: 06/22/12 Time: 20:54Sample: 199

23、8 2010Included observations: 13CoefficientStd. Errort-StatisticProb.C-17.3852015.98576-1.0875430.3023X141.715247.1387595.8434860.0002X5-0.0677320.026402-2.5654450.0281R-squared0.974402Mean dependent var273.2423Adjusted R-squared0.969282S.D. dependent var119.3226S.E. of regression20.91319Akaike info

24、criterion9.117811Sum squared resid4373.615Schwarz criterion9.248184Log likelihood-56.26577Hannan-Quinn criter.9.091014F-statistic190.3241Durbin-Watson stat2.052792Prob(F-statistic)0.000000比较可得,当加入X2时方程的R2改进最大,而且个参数的t检验显著,因此选择保留X2,再继续加入其他新变量逐步回归。Dependent Variable: YMethod: Least SquaresDate: 06/22/1

25、2 Time: 21:03Sample: 1998 2010Included observations: 13CoefficientStd. Errort-StatisticProb.C-83.9768764.45099-1.3029570.2249X116.152133.4155694.7289720.0011X20.0195020.0049403.9475290.0034X3-0.0965660.180221-0.5358190.6051R-squared0.985007Mean dependent var273.2423Adjusted R-squared0.980009S.D. dep

26、endent var119.3226S.E. of regression16.87092Akaike info criterion8.736720Sum squared resid2561.652Schwarz criterion8.910551Log likelihood-52.78868Hannan-Quinn criter.8.700990F-statistic197.0910Durbin-Watson stat1.563059Prob(F-statistic)0.000000在加入x1、x2 基础上加入x 3,但是x 3的系数符号的经济意义不符合实际,所以去除x3Dependent V

27、ariable: YMethod: Least SquaresDate: 06/22/12 Time: 21:10Sample: 1998 2010Included observations: 13CoefficientStd. Errort-StatisticProb.C-119.125526.45923-4.5022280.0015X130.0639011.083252.7125520.0239X20.0150330.0056502.6608820.0260X4-0.0137270.009800-1.4007030.1948R-squared0.987298Mean dependent v

28、ar273.2423Adjusted R-squared0.983063S.D. dependent var119.3226S.E. of regression15.52869Akaike info criterion8.570915Sum squared resid2170.261Schwarz criterion8.744745Log likelihood-51.71095Hannan-Quinn criter.8.535185F-statistic233.1760Durbin-Watson stat2.078543Prob(F-statistic)0.000000在加入x1、x2 基础上

29、加入x 4,但是x 4的系数符号的经济意义不符合实际,所以去除x4Dependent Variable: YMethod: Least SquaresDate: 06/22/12 Time: 21:12Sample: 1998 2010Included observations: 13CoefficientStd. Errort-StatisticProb.C-126.976846.33626-2.7403340.0228X110.2663114.023370.7320850.4827X20.0222840.0090422.4644350.0359X50.0130120.0391910.332

30、0230.7475R-squared0.984716Mean dependent var273.2423Adjusted R-squared0.979621S.D. dependent var119.3226S.E. of regression17.03390Akaike info criterion8.755948Sum squared resid2611.383Schwarz criterion8.929778Log likelihood-52.91366Hannan-Quinn criter.8.720218F-statistic193.2804Durbin-Watson stat1.3

31、85382Prob(F-statistic)0.000000因为加入x5后X1、x2的t检验量变的不显著,所以去除x5,所以最优组合是x1、x2最后修正严重多重共线性影响后的回归结果为:Yt=-114.9347+14.85211X1+0.019774X2t=(-4.175652)(6.410666)(4.175537)R2=0.984529 R2=0.981434 F=318.1755 DW=1.4156964.异方差检验及其修正异方差检验1).图像法从图上看,无法准确判断是否存在异方差,下面我们运用WHITE检验进一步检验模型的异方差是否存在。White检验Heteroskedasticit

32、y Test: WhiteF-statistic0.641701Prob. F(5,7)0.6768Obs*R-squared4.085866Prob. Chi-Square(5)0.5371Scaled explained SS2.965640Prob. Chi-Square(5)0.7053Test Equation:Dependent Variable: RESID2Method: Least SquaresDate: 06/22/12 Time: 21:42Sample: 1998 2010Included observations: 13CoefficientStd. Errort-

33、StatisticProb.C-381.06253848.275-0.0990220.9239X1-349.0453894.9700-0.3900080.7081X12-39.7948035.11834-1.1331630.2945X1*X20.1297690.1419920.9139200.3912X20.4864591.2933400.3761260.7180X22-9.83E-050.000138-0.7111940.5000R-squared0.314297Mean dependent var203.3361Adjusted R-squared-0.175490S.D. depende

34、nt var331.4905S.E. of regression359.4021Akaike info criterion14.91080Sum squared resid904189.2Schwarz criterion15.17154Log likelihood-90.92019Hannan-Quinn criter.14.85720F-statistic0.641701Durbin-Watson stat3.332954Prob(F-statistic)0.676836从上表看出,在=0.05的显著水平下,P值大于,所以模型不存在异方差5.自相关的检验及其修正因为散点图大致呈递增趋势,所以需进行进一步检验因为折线图没有连续为正或连续为负的现象,所以需进行SCLM检验:Breusch-Godfrey Serial Correlation LM Test:F-statistic0.630224Prob. F(1,9)0.4477Obs*R-squared0.850750Prob. Chi-Square(1)0.3563Test Equa

温馨提示

  • 1. 本站所有资源如无特殊说明,都需要本地电脑安装OFFICE2007和PDF阅读器。图纸软件为CAD,CAXA,PROE,UG,SolidWorks等.压缩文件请下载最新的WinRAR软件解压。
  • 2. 本站的文档不包含任何第三方提供的附件图纸等,如果需要附件,请联系上传者。文件的所有权益归上传用户所有。
  • 3. 本站RAR压缩包中若带图纸,网页内容里面会有图纸预览,若没有图纸预览就没有图纸。
  • 4. 未经权益所有人同意不得将文件中的内容挪作商业或盈利用途。
  • 5. 人人文库网仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对用户上传分享的文档内容本身不做任何修改或编辑,并不能对任何下载内容负责。
  • 6. 下载文件中如有侵权或不适当内容,请与我们联系,我们立即纠正。
  • 7. 本站不保证下载资源的准确性、安全性和完整性, 同时也不承担用户因使用这些下载资源对自己和他人造成任何形式的伤害或损失。

评论

0/150

提交评论