我国涉外旅游业收入的实证分析.docx

1、我国涉外旅游业收入的实证分析我国涉外旅游业收入的实证分析一概况分析涉外旅游是指我国旅游部门经营对外招徕并接待外国人、外籍华人和华侨等国际旅行者旅游业务的活动。涉外旅游业是一项新型的综合性的经济事业。旅游业是日益兴起的新型“朝阳产业”，它的发展无论是对一国的经济，还是国际间的交流，都有着重要和积极的作用。中国是世界上旅游资源和市场最丰富的国家，把握这一优势，大力发展旅游业，对正在深入的改革开放和产业结构的优化，都有着广泛的促进作用。近年来，我国旅游业突飞猛进。随着我国对外开放的逐步深入，涉外旅游业也获得了长足的发展。它是我国国民经济和发展对外经济关系的一个重要组成部分，是第三产业的重要部门。 中

2、国旅游市场在21世纪将进一步扩大，中国丰富的旅游资源不断得到开发；旅游产品结构不断完善；旅游产业规模不断扩大；发展旅游的大环境逐渐优化，这些都为中国旅游市场的扩大提供了坚实的保障。我国涉外旅游市场将会继续扩大，亚洲是中国的最大客源市场，随着东南亚金融危机的过去，东南亚、日本的经济复苏，亚洲客源肯定有较大的发展；欧美远程客源国来华人数都在不断增长，在中国国际旅游市场上，来自欧美的游客只是一个全球的平均水平，欧美来华旅游的潜力显然很大。二 模型的建立我们通过分析我国涉外旅游业的收入，根据理论及对现实情况的认识,建立了一个单一方程模型: Y=1+2X2+3X3+4X4+5X5+U (1.1) 其中：

3、Y我国涉外旅游业收入（亿元）X2涉外饭店数目（个）X3旅游人数（万人）X4涉外旅游业职工人数（人）X5涉外旅行社个数（个）U 随及扰动项 i参数三模型的估计和检验（一）估计设模型中的随及误差项U满足古典假定，运用OLS方法估计未知参数，利用计量经济学计算机软件Eviews计算的过程如下：1建立文档，输入数据首先点击Eviews图标，进入Eviews主页。建立新的Workfile工作框，并输入数据，见表一。表一obsX2X3X4X5Y1991 2130.000 3335.000 38177.00 671.0000 28.400001992 2354.000 3811.500 40258.00 8

4、52.0000 39.500001993 2552.000 4152.700 45431.00 987.0000 46.800001994 2995.000 4368.450 57600.00 1110.000 73.230001995 3720.000 4638.650 59935.00 1025.000 87.330001996 4418.000 5112.750 53093.00 977.0000 102.00001997 5201.000 5758.790 48881.00 991.0000 120.74001998 5782.000 6347.840 52290.00 1312.00

5、0 126.02001999 7035.000 7279.560 47153.00 1256.000 140.99002OLS估计未知参数在主页上选Group菜单，点击Estimate Equation项，对数据进行OLS估计，结果如表二表二Dependent Variable: YMethod: Least SquaresDate: 05/15/04 Time: 10:29Sample: 1991 1999Included observations: 9VariableCoefficientStd. Errort-StatisticProb. X20.0108310.0197650.5480

6、080.6128X30.0192740.0312600.6165770.5709X40.0016520.0005592.9537510.0418X5-0.0464840.050411-0.9221040.4087C-88.3649852.76179-1.6747910.1693R-squared0.988138 Mean dependent var85.00111Adjusted R-squared0.976277 S.D. dependent var40.71376S.E. of regression6.270853 Akaike info criterion6.809883Sum squa

7、red resid157.2944 Schwarz criterion6.919452Log likelihood-25.64447 F-statistic83.30613Durbin-Watson stat1.852069 Prob(F-statistic)0.000419（二）检验1经济意义检验X5的系数与其经济意义不符。我们将通过对模型的修正看是否能得到更好的结果。2 统计检验对回归系数进行整体检验，该检验是在方差分析的基础上利用F检验进行的。由上表数据，F=248.8175F0.05（4，4），应该拒绝原假设H0，说明回归方程显著。所以从模型从整体上看，涉外旅游收入与解释变量之间线形关

8、系显著3 计量经济学检验（1）.多重共线性检验在Quick菜单中选取项Group Statistics中的Correlation命令，输入变量名即可得到如下结果：表三 X2X3 X4 X5X2 1.000000 0.9918410.297759 0.777257X30.991841 1.0000000.3029190.832983 X40.297759 0.3029191.000000 0.573134X50.7772570.8329830.5731341.000000由表3可以看出，解释变量之间存在高度线性相关。同时由表2得到的可决系数很大，而且F统计量值显著的大于给定显著性水平下的临界值，

9、而x2、x3、x5变量的偏回归系数t统计量值并不显著，且X5的系数的符号与经济意义相悖。尽管整体上线性回归拟合较好，但模型中解释变量存在严重的多重共线性。对多重共线性的修正a）.运用OLS方法逐一求对各个解释变量的回归。结合经济意义和统计检验选出拟合效果最好的一元线形回归方程。Y对X2的回归表四Dependent Variable: YMethod: Least SquaresDate: 05/15/04 Time: 14:56Sample: 1991 1999Included observations: 9VariableCoefficientStd. Errort-StatisticPro

10、b. X20.0230980.00218510.571100.0000C-7.8688979.465154-0.8313540.4332R-squared0.941052 Mean dependent var85.00111Adjusted R-squared0.932630 S.D. dependent var40.71376S.E. of regression10.56752 Akaike info criterion7.746577Sum squared resid781.7067 Schwarz criterion7.790404Log likelihood-32.85959 F-st

11、atistic111.7481Durbin-Watson stat0.736732 Prob(F-statistic)0.000015将上述回归结果整理如下：Y=-7.868897+0.023098X2 （1.2） （-0.8313）（10.5711）R2=0.9410 ，S.E =10.5675， F=111.7481Y对X3的回归表五Dependent Variable: YMethod: Least SquaresDate: 05/15/04 Time: 15:06Sample: 1991 1999Included observations: 9VariableCoefficientSt

12、d. Errort-StatisticProb. X30.0306410.0033589.1260270.0000C-67.5394117.19614-3.9275900.0057R-squared0.922467 Mean dependent var85.00111Adjusted R-squared0.911391 S.D. dependent var40.71376S.E. of regression12.11937 Akaike info criterion8.020616Sum squared resid1028.153 Schwarz criterion8.064444Log li

13、kelihood-34.09277 F-statistic83.28438Durbin-Watson stat0.775610 Prob(F-statistic)0.000039将上述回归结果整理如下：Y= -67.53941+0.030641X3 （1.3）（-3.9275）（9.1260） R2= 0.9224， S.E=12.1193， F=83.2843Y对X4的回归表六Dependent Variable: YMethod: Least SquaresDate: 05/15/04 Time: 15:07Sample: 1991 1999Included observations: 9

14、VariableCoefficientStd. Errort-StatisticProb. X40.0027130.0018301.4823250.1818C-48.4778090.93217-0.5331200.6105R-squared0.238906 Mean dependent var85.00111Adjusted R-squared0.130178 S.D. dependent var40.71376S.E. of regression37.97137 Akaike info criterion10.30467Sum squared resid10092.77 Schwarz cr

15、iterion10.34850Log likelihood-44.37102 F-statistic2.197288Durbin-Watson stat0.301641 Prob(F-statistic)0.181813将上述回归结果整理如下：Y=-48.4778+0.0027X4 （1.4）（-0.5331） （1.4823） R2=0.2389，S.E=37.9713，F=2.1972Y对X5的回归表七Dependent Variable: YMethod: Least SquaresDate: 05/15/04 Time: 15:08Sample: 1991 1999Included o

16、bservations: 9VariableCoefficientStd. Errort-StatisticProb. X50.1684920.0468923.5932230.0088C-86.8795048.60200-1.7875700.1170R-squared0.648440 Mean dependent var85.00111Adjusted R-squared0.598217 S.D. dependent var40.71376S.E. of regression6 Akaike info criterion9.532296Sum squared resid4661.995 Sch

17、warz criterion9.576123Log likelihood-40.89533 F-statistic12.91125Durbin-Watson stat1.119498 Prob(F-statistic)0.008819将上述回归结果整理如下： Y=-86.8795+0.1684X5 （1.5） （-1.7875）（3.5932） R2 =0.6484，S.E=25.8069， F=12.9112分析：通过对多重可决系数和t统计量的观察，X2，X3的可决系数接近1，且t的绝对值都远大于2，所以模型对数据的拟合程度较好。同时结合经济意义，旅游人数x3较涉外饭店数x2对涉外旅游收入y

18、的影响更大，选出x3，得一元线性回归方程：Y=-67.53941+0.030641X3 b. 逐步回归。引入其余解释变量，得到以下模型。表八Dependent Variable: YMethod: Least SquaresDate: 05/16/04 Time: 00:02Sample: 1991 1999Included observations: 9VariableCoefficientStd. Errort-StatisticProb. X3-0.0033610.024767-0.1357050.8965X20.0255860.0184851.3841330.2156C-1.14010

19、250.62397-0.0225210.9828R-squared0.941232 Mean dependent var85.00111Adjusted R-squared0.921643 S.D. dependent var40.71376S.E. of regression11.39675 Akaike info criterion7.965734Sum squared resid779.3148 Schwarz criterion8.031476Log likelihood-32.84580 F-statistic48.04823Durbin-Watson stat0.742719 Pr

20、ob(F-statistic)0.000203表九Dependent Variable: YMethod: Least SquaresDate: 05/16/04 Time: 00:03Sample: 1991 1999Included observations: 9VariableCoefficientStd. Errort-StatisticProb. X30.0285360.00253611.251910.0000X40.0012090.0004412.7401750.0337C-116.545721.75047-5.3583060.0017R-squared0.965563 Mean

21、dependent var85.00111Adjusted R-squared0.954084 S.D. dependent var40.71376S.E. of regression8.724176 Akaike info criterion7.431275Sum squared resid456.6675 Schwarz criterion7.497016Log likelihood-30.44074 F-statistic84.11512Durbin-Watson stat1.174184 Prob(F-statistic)0.000041表十Dependent Variable: YM

22、ethod: Least SquaresDate: 05/16/04 Time: 00:05Sample: 1991 1999Included observations: 9VariableCoefficientStd. Errort-StatisticProb. X30.0301880.0065514.6083920.0037X50.0035660.0429640.0830040.9365C-68.9225124.94512-2.7629660.0327R-squared0.922556 Mean dependent var85.00111Adjusted R-squared0.896742

23、 S.D. dependent var40.71376S.E. of regression13.08290 Akaike info criterion8.241691Sum squared resid1026.974 Schwarz criterion8.307432Log likelihood-34.08761 F-statistic35.73773Durbin-Watson stat0.780006 Prob(F-statistic)0.000464Y=-1.140102+0.025586X2+-0.003361X3 （-0.0225） （1.3841） （-0.1357）Adjusted

24、 R2=0.9216 S.E=11.3967 F=48.0482Y=-116.5457+0.028536X3+0.001209X4 (-5.3583) (11.2519) (2.7401)Adjusted R2=0.9540 S.E= 8.7241 F=84.1151Y=-68.92251+0.030188X3+0.003566X5 (-2.7629) (4.6083) (0.0830) Adjusted R2=0.8967 S.E=13.0829 F=35.7377可以看出X2和X5的对模型的拟合优度并无改善，同时对X3影响较小。X4在符合经济意义的前提下，使拟合优度提高，每个参数统计检验显

25、著，应采纳该变量。得到一个二元回归方程： Y=-116.5457+0.028536X3+0.001209X4表十一Dependent Variable: YMethod: Least SquaresDate: 05/16/04 Time: 22:03Sample: 1991 1999Included observations: 9VariableCoefficientStd. Errort-StatisticProb. X3-0.0066550.013448-0.4948780.6417X40.0012270.0003123.9280740.0111X20.0264560.0100202.64

26、03700.0460C-48.6260629.98005-1.6219470.1657R-squared0.985617 Mean dependent var85.00111Adjusted R-squared0.976987 S.D. dependent var40.71376S.E. of regression6.176249 Akaike info criterion6.780402Sum squared resid190.7303 Schwarz criterion6.868057Log likelihood-26.51181 F-statistic114.2115Durbin-Wat

27、son stat1.743783 Prob(F-statistic)0.000050表十二Dependent Variable: YMethod: Least SquaresDate: 05/16/04 Time: 22:09Sample: 1991 1999Included observations: 9VariableCoefficientStd. Errort-StatisticProb. X30.0363030.00315511.506970.0001X40.0018660.0003715.0363880.0040X5-0.0701690.024064-2.9159190.0332C-

28、115.980114.50022-7.9985070.0005R-squared0.987248 Mean dependent var85.00111Adjusted R-squared0.979597 S.D. dependent var40.71376S.E. of regression5.815561 Akaike info criterion6.660054Sum squared resid169.1038 Schwarz criterion6.747709Log likelihood-25.97024 F-statistic129.0310Durbin-Watson stat1.949370 Prob(F-statistic)0.000037分析： 由表11、表12知，分别引入x2或x5后，他们对y的影响并不显著，故将x2和x5删除，此后统计检验效果均有较大改善。综上所述，选择此模型为修正后的