中國民航業(yè)主營業(yè)務(wù)收入影響因素分析

1、中國民航業(yè)主營業(yè)務(wù)收入影響因素分析計(jì)量經(jīng)濟(jì)學(xué)論文中國民航業(yè)主營業(yè)務(wù)收入影響因素分析班級：09><>8統(tǒng)計(jì)三班組員：程馳（40<>809147） 李瀚（40<>80914<>8）目錄：研究背景與研究意義變量的選取模型的建立與分析（一）變量平穩(wěn)性檢驗(yàn)（二）設(shè)定誤差檢驗(yàn)（RESET檢驗(yàn)）（三）多重共線性檢驗(yàn)及處理（四）遺漏及冗余變量檢驗(yàn)（五）協(xié)整檢驗(yàn)（六）長期與短期模型四、結(jié)論與建議五、附件（一）相關(guān)圖表（二）原始數(shù)據(jù)參考文獻(xiàn)研究背景與研究意義中國民航是國民經(jīng)濟(jì)的重要組成部分， 是國家重要的基礎(chǔ)性、 先導(dǎo)性產(chǎn)業(yè)， 也是推動經(jīng)濟(jì)社會發(fā)展的戰(zhàn)略性和創(chuàng)

2、新性產(chǎn)業(yè)。中國民航運(yùn)輸總周轉(zhuǎn)量已經(jīng)連續(xù)數(shù)年位居世界第二位，是名副其實(shí)的航空大國。從民航大國逐步向民航強(qiáng)國轉(zhuǎn)變，是今后我國民航的戰(zhàn)略目標(biāo)。雖然我國航空業(yè)已經(jīng)逐漸發(fā)展壯大，但在當(dāng)前國內(nèi)外錯(cuò)綜復(fù)雜的經(jīng)濟(jì)環(huán)境中仍然是危機(jī)四伏，特別是剛剛過去的全球性金融危機(jī)以及全球石油價(jià)格的上漲，更是對中國航空業(yè)造成了巨大的沖擊與影響。許多航空公司營業(yè)收入大幅下降，甚至面臨倒閉的危險(xiǎn)。我們小組希望對民航業(yè)運(yùn)輸總收入的研究找出影響中國航空業(yè)發(fā)展的因素，為加快中國航空的發(fā)展找出相應(yīng)合理并且有針對性的方針對策做準(zhǔn)備。從而使中國民航業(yè)在復(fù)雜而且危機(jī)四伏的經(jīng)濟(jì)環(huán)境中能夠克服各種困難穩(wěn)步發(fā)展。變量的選取經(jīng)過篩選，最后我們找出了四個(gè)

3、與中國民航主營業(yè)務(wù)相關(guān)的變量，分別是： X2-民航客運(yùn)周轉(zhuǎn)量客運(yùn)周轉(zhuǎn)量是指在一定時(shí)期內(nèi)運(yùn)送旅客數(shù)量與平均運(yùn)距的乘積，計(jì)量單位是“人公里”。民航客運(yùn)周轉(zhuǎn)量反映了一定時(shí)期內(nèi)民航總運(yùn)輸旅客數(shù)量與平均運(yùn)送距離的乘積。X3-民航貨運(yùn)周轉(zhuǎn)量貨運(yùn)周轉(zhuǎn)量是指運(yùn)輸貨物的數(shù)量（噸）與運(yùn)輸距離（公里）的乘積。其表示方法為噸公里。而且，民航主要由客運(yùn)和貨運(yùn)量部分組成，這兩個(gè)變量可以較好的反映民航整體收入情況。X4-民航業(yè)航空燃油消耗量民航業(yè)航空燃油消耗量，客觀的反映了民航業(yè)成本消費(fèi)，相比國際油價(jià)變動來說，燃油消耗量變動更為有規(guī)律，對用年份數(shù)據(jù)的分析更為方便。而且關(guān)于燃油附加稅的調(diào)整，對國際油價(jià)變動有一定抵消的作用，如

4、果在使用國際油價(jià)變動作為變量，可能會出現(xiàn)較大的偏差。X5-城鎮(zhèn)家庭平均每人全年交通和通信消費(fèi)性支出城鎮(zhèn)家庭平均每人全年交通和通信消費(fèi)性支出，在一定方面反映了民航客運(yùn)的購買力。我們只加入了城鎮(zhèn)家庭消費(fèi)，是因?yàn)槌擎?zhèn)人均對民航消費(fèi)的能力和頻率遠(yuǎn)遠(yuǎn)大于農(nóng)村人均對民航的消費(fèi)能力。被解釋變量Y-民航企業(yè)運(yùn)輸收入合計(jì)模型的建立與分析（一）變量平穩(wěn)性檢驗(yàn)對Y X2,X3,X4,X5.分別做單位根檢驗(yàn)其平穩(wěn)性結(jié)果如下：N-不能拒絕H0Y-拒絕H0表1Y含截距項(xiàng)含截距項(xiàng)與趨勢項(xiàng)無LevelNNN一階差分NNN這里Y的ADF檢驗(yàn)無法滿足一階差分條件，我們通過觀察其散點(diǎn)圖發(fā)現(xiàn)其為非光滑曲線，故試用PP檢驗(yàn)，結(jié)果滿足一

5、階差分條件。表2Null Hypothesis: D(Y) has a unit rootExogenous: NoneBandwidth: 3 (Newey-West using Bartlett kernel)Adj. t-Stat?Prob.*Phillips-Perron test statistic-2.160262?0.0332Test critical values:1% level-2.70<>80945% level-1.962<>81310% level-1.606129*MacKinnon (1996) one-sided p-values.War

6、ning: Probabilities and critical values calculated for 20 observations?and may not be accurate for a sample size of 17Residual variance (no correction)?1.79E+12HAC corrected variance (Bartlett kernel)?1.90E+12表3X2含截距項(xiàng)含截距項(xiàng)與趨勢項(xiàng)無LevelNNN一階差分NYN表4表5X3含截距項(xiàng)含截距項(xiàng)與趨勢項(xiàng)無LevelNNN一階差分YYN表6表7表<>8X4含截距項(xiàng)含截距項(xiàng)與

7、趨勢項(xiàng)無LevelNNN一階差分YYN表9表10表11X5含截距項(xiàng)含截距項(xiàng)與趨勢項(xiàng)無LevelNNN一階差分NYN表12 由上圖可知，在5%的顯著水平下各變量序列均為非平穩(wěn)的，Y,X2,X3,X4,X5序列都是一階單整的。 以上變量均是一階單整的，可以構(gòu)建計(jì)量經(jīng)濟(jì)學(xué)模型。（二）RESET檢驗(yàn)表13Ramsey RESET Test:F-statistic0.406<>8<>84?Prob. F(3,11)0.7511Log likelihood ratio1.999409?Prob. Chi-Square(3)0.5725Test Equation:Dependent

8、Variable: YMethod: Least SquaresDate: 12/10/10 Time: 16:55Sample: 1990 200<>8Included observations: 19VariableCoefficientStd. Errort-StatisticProb.?C-29<>8330.1554339.0-0.53<>81730.6012X21044<>8.933435.2613.0416720.0112X31692.3<>8334036.350.0497230.9612X4-1.4436670.6<

9、;>86716-2.1022770.0594X54939.71546<>87.1<>841.053<>8770.3145FITTED21.72E-0<>89.54E-0<>80.1<>801650.<>8603FITTED3-1.70E-157.37E-15-0.2307350.<>821<>8FITTED44.0<>8E-231.79E-220.22<>835<>80.<>8236R-squared0.997952?Mean dep

10、endent var7293549.Adjusted R-squared0.996649?S.D. dependent var60054<>8<>8.S.E. of regression347645.2?Akaike info criterion2<>8.65131Sum squared resid1.33E+12?Schwarz criterion29.04<>897Log likelihood-264.1<>875?Hannan-Quinn criter.2<>8.71<>861F-statistic765

11、.7<>87<>8?Durbin-Watson stat2.97195<>8Prob(F-statistic)0.000000未落入拒絕域中，不存在設(shè)定誤差。（三）多重共線性檢驗(yàn)及處理逐步回歸法：分別做Y對X2,X3,X4,X5的一元回歸。結(jié)果如下：表14Dependent Variable: YMethod: Least SquaresDate: 12/17/10 Time: 0<>8:34Sample: 1990 200<>8Included observations: 19VariableCoefficientStd. Erro

12、rt-StatisticProb.?C-12<>83723.15675<>8.2-<>8.1<>891<>8<>80.0000X27310.134110.46<>8766.173<>810.0000R-squared0.996133?Mean dependent var7293549.Adjusted R-squared0.995905?S.D. dependent var60054<>8<>8.S.E. of regression3<>842<>8&

13、lt;>8.2?Akaike info criterion2<>8.65547Sum squared resid2.51E+12?Schwarz criterion2<>8.754<>89Log likelihood-270.2270?Hannan-Quinn criter.2<>8.67230F-statistic437<>8.973?Durbin-Watson stat1.612497Prob(F-statistic)0.000000表15Dependent Variable: YMethod: Least SquaresD

14、ate: 12/17/10 Time: 0<>8:35Sample: 1990 200<>8Included observations: 19VariableCoefficientStd. Errort-StatisticProb.?C-<>8764<>87.<>83<>80939.<>8-2.300<>8560.0343X3171<>826.36542.42<>826.2633<>80.0000R-squared0.975947?Mean dependent v

15、ar7293549.Adjusted R-squared0.974532?S.D. dependent var60054<>8<>8.S.E. of regression95<>8400.6?Akaike info criterion30.4<>8322Sum squared resid1.56E+13?Schwarz criterion30.5<>8263Log likelihood-2<>87.5906?Hannan-Quinn criter.30.50004F-statistic6<>89.7649?Du

16、rbin-Watson stat0.<>822993Prob(F-statistic)0.000000表16Dependent Variable: YMethod: Least SquaresDate: 12/17/10 Time: 0<>8:35Sample: 1990 200<>8Included observations: 19VariableCoefficientStd. Errort-StatisticProb.?C-1753036.3<>82743.0-4.5<>801920.0003X41.764<>8910

17、.0629122<>8.053100.0000R-squared0.97<>8<>855?Mean dependent var7293549.Adjusted R-squared0.977611?S.D. dependent var60054<>8<>8.S.E. of regression<>89<>8592.7?Akaike info criterion30.35435Sum squared resid1.37E+13?Schwarz criterion30.45376Log likelihood-2<

18、;>86.3663?Hannan-Quinn criter.30.37117F-statistic7<>86.9765?Durbin-Watson stat0.602712Prob(F-statistic)0.000000表17Dependent Variable: YMethod: Least SquaresDate: 12/17/10 Time: 0<>8:35Sample: 1990 200<>8Included observations: 19VariableCoefficientStd. Errort-StatisticProb.?C7207

19、65.9302515.72.3<>825740.0291X513050.9<>8450.<>839<>82<>8.94<>8150.0000R-squared0.9<>80117?Mean dependent var7293549.Adjusted R-squared0.97<>8947?S.D. dependent var60054<>8<>8.S.E. of regression<>871370.1?Akaike info criterion30.292<

20、;>82Sum squared resid1.29E+13?Schwarz criterion30.39224Log likelihood-2<>85.7<>81<>8?Hannan-Quinn criter.30.30965F-statistic<>837.9955?Durbin-Watson stat0.<>8319<>80Prob(F-statistic)0.000000綜上，各自回歸結(jié)果如下表：表1<>8變量X2X3X4X5參數(shù)估計(jì)值7310.134171<>826.31.764<

21、;>89113050.9<>8t66.173<>8126.2633<>82<>8.053102<>8.94<>815R20.9961330.9759470.97<>8<>8550.9<>80117adjust R20.9959050.9745320.9776110.97<>8947可見，X2的adjust R2最大，且t值也顯著。故保留。分別做Y與X2對X3,X4,X5的回歸。結(jié)果如下：表19Dependent Variable: YMethod: Least Squa

22、resDate: 12/0<>8/10 Time: 23:01Sample: 1990 200<>8Included observations: 19VariableCoefficientStd. Errort-StatisticProb.?C-131<>89<>83.15<>8024.1-<>8.3467210.0000X2<>8314.042<>866.<>82159.5914120.0000X3-24031.69205<>84.46-1.16746<>80.

23、2601R-squared0.996436?Mean dependent var7293549.Adjusted R-squared0.995991?S.D. dependent var60054<>8<>8.S.E. of regression3<>80250.1?Akaike info criterion2<>8.67<>899Sum squared resid2.31E+12?Schwarz criterion2<>8.<>82<>811Log likelihood-269.4504?Hann

24、an-Quinn criter.2<>8.70422F-statistic2236.91<>8?Durbin-Watson stat1.953715Prob(F-statistic)0.000000表20Dependent Variable: YMethod: Least SquaresDate: 12/0<>8/10 Time: 23:01Sample: 1990 200<>8Included observations: 19VariableCoefficientStd. Errort-StatisticProb.?C-1122047.1630

25、66.3-6.<>8<>8092<>80.0000X29277.54<>8951.12069.7543340.0000X4-0.4<>81<>8940.231647-2.0<>802<>890.0539R-squared0.996956?Mean dependent var7293549.Adjusted R-squared0.996576?S.D. dependent var60054<>8<>8.S.E. of regression351429.7?Akaike info

26、 criterion2<>8.52135Sum squared resid1.9<>8E+12?Schwarz criterion2<>8.67047Log likelihood-267.952<>8?Hannan-Quinn criter.2<>8.5465<>8F-statistic2620.224?Durbin-Watson stat2.012<>843Prob(F-statistic)0.000000表21Dependent Variable: YMethod: Least SquaresDate: 1

27、2/0<>8/10 Time: 23:01Sample: 1990 200<>8Included observations: 19VariableCoefficientStd. Errort-StatisticProb.?C-135<>8210.2<>89033.7-4.6991420.0002X27596.564929.7373<>8.1706560.0000X5-519.41<>891673.392-0.3103990.7603R-squared0.996156?Mean dependent var7293549.Ad

28、justed R-squared0.995675?S.D. dependent var60054<>8<>8.S.E. of regression394927.9?Akaike info criterion2<>8.75473Sum squared resid2.50E+12?Schwarz criterion2<>8.903<>86Log likelihood-270.1700?Hannan-Quinn criter.2<>8.77997F-statistic2073.150?Durbin-Watson stat1.65

29、6160Prob(F-statistic)0.000000由上圖可見，加入X3,X4,X5變量后模型出現(xiàn)了共線性造成的問題。故剔除這三個(gè)變量。并對被剔除的三個(gè)變量進(jìn)行遺漏變量檢驗(yàn)。（四）遺漏與冗余變量檢驗(yàn)1.對X3進(jìn)行遺漏變量檢驗(yàn)表22Omitted Variables: X3F-statistic1.3629<>81?Prob. F(1,16)0.2601Log likelihood ratio1.5532<>82?Prob. Chi-Square(1)0.2127Test Equation:Dependent Variable: YMethod: Least Squ

30、aresDate: 12/21/10 Time: 13:39Sample: 1990 200<>8Included observations: 19VariableCoefficientStd. Errort-StatisticProb.?C-131<>89<>83.15<>8024.1-<>8.3467210.0000X2<>8314.042<>866.<>82159.5914120.0000X3-24031.69205<>84.46-1.16746<>80.2601R-s

31、quared0.996436?Mean dependent var7293549.Adjusted R-squared0.995991?S.D. dependent var60054<>8<>8.S.E. of regression3<>80250.1?Akaike info criterion2<>8.67<>899Sum squared resid2.31E+12?Schwarz criterion2<>8.<>82<>811Log likelihood-269.4504?Hannan-Quin

32、n criter.2<>8.70422F-statistic2236.91<>8?Durbin-Watson stat1.953715Prob(F-statistic)0.000000對X4進(jìn)行遺漏變量檢驗(yàn)表23Omitted Variables: X4F-statistic4.327604?Prob. F(1,16)0.0539Log likelihood ratio4.54<>8430?Prob. Chi-Square(1)0.0329Test Equation:Dependent Variable: YMethod: Least SquaresDate

33、: 12/21/10 Time: 13:43Sample: 1990 200<>8Included observations: 19VariableCoefficientStd. Errort-StatisticProb.?C-1122047.163066.3-6.<>8<>8092<>80.0000X29277.54<>8951.12069.7543340.0000X4-0.4<>81<>8940.231647-2.0<>802<>890.0539R-squared0.996956?M

34、ean dependent var7293549.Adjusted R-squared0.996576?S.D. dependent var60054<>8<>8.S.E. of regression351429.7?Akaike info criterion2<>8.52135Sum squared resid1.9<>8E+12?Schwarz criterion2<>8.67047Log likelihood-267.952<>8?Hannan-Quinn criter.2<>8.5465<>

35、8F-statistic2620.224?Durbin-Watson stat2.012<>843Prob(F-statistic)0.000000對X5進(jìn)行遺漏變量檢驗(yàn)表24Omitted Variables: X5F-statistic0.096347?Prob. F(1,16)0.7603Log likelihood ratio0.114069?Prob. Chi-Square(1)0.7356Test Equation:Dependent Variable: YMethod: Least SquaresDate: 12/21/10 Time: 13:43Sample: 19

36、90 200<>8Included observations: 19VariableCoefficientStd. Errort-StatisticProb.?C-135<>8210.2<>89033.7-4.6991420.0002X27596.564929.7373<>8.1706560.0000X5-519.41<>891673.392-0.3103990.7603R-squared0.996156?Mean dependent var7293549.Adjusted R-squared0.995675?S.D. depende

37、nt var60054<>8<>8.S.E. of regression394927.9?Akaike info criterion2<>8.75473Sum squared resid2.50E+12?Schwarz criterion2<>8.903<>86Log likelihood-270.1700?Hannan-Quinn criter.2<>8.77997F-statistic2073.150?Durbin-Watson stat1.656160Prob(F-statistic)0.0000005.再對X2進(jìn)行

38、冗余變量檢驗(yàn) 表25Redundant Variables: X2F-statistic437<>8.973?Prob. F(1,17)0.0000Log likelihood ratio105.5494?Prob. Chi-Square(1)0.0000Test Equation:Dependent Variable: YMethod: Least SquaresDate: 12/21/10 Time: 14:36Sample: 1990 200<>8Included observations: 19VariableCoefficientStd. Errort-Sta

39、tisticProb.?C7293549.1377753.5.29379<>80.0000R-squared0.000000?Mean dependent var7293549.Adjusted R-squared0.000000?S.D. dependent var60054<>8<>8.S.E. of regression60054<>8<>8.?Akaike info criterion34.10544Sum squared resid6.49E+14?Schwarz criterion34.15515Log likelihoo

40、d-323.0017?Hannan-Quinn criter.34.113<>85Durbin-Watson stat0.067279 （五）協(xié)整檢驗(yàn)?zāi)Ｐ突貧w結(jié)果如下：表26Dependent Variable: YMethod: Least SquaresDate: 12/17/10 Time: 0<>8:47Sample: 1990 200<>8Included observations: 19VariableCoefficientStd. Errort-StatisticProb.?C-12<>83723.15675<>8.2-

41、<>8.1<>891<>8<>80.0000X27310.134110.46<>8766.173<>810.0000R-squared0.996133?Mean dependent var7293549.Adjusted R-squared0.995905?S.D. dependent var60054<>8<>8.S.E. of regression3<>842<>8<>8.2?Akaike info criterion2<>8.65547Sum s

42、quared resid2.51E+12?Schwarz criterion2<>8.754<>89Log likelihood-270.2270?Hannan-Quinn criter.2<>8.67230F-statistic437<>8.973?Durbin-Watson stat1.612497Prob(F-statistic)0.0000002. 對該模型殘差序列做單位根檢驗(yàn)結(jié)果如下：注：因?yàn)槭菃我唤忉屪兞浚薀o需用C（）查表求。含截距：表27Null Hypothesis: EE has a unit rootExogenous: ConstantLag Lengt