《對我國財政收入影響因素的回歸分析 計量經(jīng)濟學(xué)論文》由會員分享，可在線閱讀，更多相關(guān)《對我國財政收入影響因素的回歸分析 計量經(jīng)濟學(xué)論文（5頁珍藏版）》請在裝配圖網(wǎng)上搜索。
1、對我國財政收入影響因素的回歸分析 摘要:本文通過引入多元回歸分析的方法來分析對我國財政收入有影響的幾個因素。 關(guān)鍵詞異方差 自相關(guān) 多重共線性 財政收入對于國民經(jīng)濟的運行及社會發(fā)展具有重要影響。首先它是一個國家各項收入得以實現(xiàn)的物質(zhì)保證。一個國家財政收入規(guī)模大小往往是衡量其經(jīng)濟實力的重要標志。其次財政收入是國家對經(jīng)濟實行宏觀調(diào)控的重要經(jīng)濟杠桿。宏觀調(diào)控的首要問題是社會總需求與總供給的平衡問題實現(xiàn)社會總需求與總供給的平衡包括總量上的平衡和結(jié)構(gòu)上的平衡兩個層次的內(nèi)容。財政收入的杠桿既可通過增收和減收來發(fā)揮總量調(diào)控作用也可通過對不同財政資金繳納者的財政負擔(dān)大小的調(diào)整來發(fā)揮結(jié)構(gòu)調(diào)整的作用。此外財政收入
2、分配也是調(diào)整國民收入初次分配格局實現(xiàn)社會財富公平合理分配的主要工具。在我國財政收入的主體是稅收收入。因此在稅收體制及政策不變的情況下財政收入會隨著經(jīng)濟繁榮而增加隨著經(jīng)濟衰退而下降。 我國的財政收入主要包括稅收、國有經(jīng)濟收入、債務(wù)收入以及其他收入四種形式因此財政收入會受到不同因素的影響。從國民經(jīng)濟部門結(jié)構(gòu)看財政收入又表現(xiàn)為來自各經(jīng)濟部門的收入。財政收入的部門構(gòu)成就是在財政收入中由來自國民經(jīng)濟各部門的收入所占的不同比例來表現(xiàn)財政收入來源的結(jié)構(gòu)它體現(xiàn)國民經(jīng)濟各部門與財政收入的關(guān)系。我國財政收入主要來自于工業(yè)、農(nóng)業(yè)、商業(yè)、交通運輸和服務(wù)業(yè)等部門。 因此本文認為財政收入主要受到第一產(chǎn)業(yè)增加值、第二產(chǎn)業(yè)增
3、加值、第三產(chǎn)業(yè)增加值、其他收入水平和就業(yè)人口總數(shù)的影響。令財政收入為Y億元、第一產(chǎn)業(yè)增加值為X1億元、第二產(chǎn)業(yè)增加值為X2億元、第三產(chǎn)業(yè)增加值為X3億元、就業(yè)人口總數(shù)為X4億人其他收入為X5億元據(jù)此建立模型方程為: ya1a2X2a3X3a4X4a5X5u 并取1978年—2002年的數(shù)據(jù)如下 表1 Y X1 X2 X3 X4 X5 1978 1132.260 1018.400 1745.200 860.5000 40152.00 40.99000 1979 1146.380 1258.900 1913.500 865.8000 41024.00 113.5300 1980 1159.930
4、1359.400 2192.000 966.4000 42361.00 152.9900 1981 1175.790 1545.600 2255.500 1061.300 43725.00 192.2200 1982 1212.330 1761.600 2383.000 1150.100 45295.00 215.8400 1983 1366.950 1960.800 2646.200 1327.500 46436.00 257.8400 1984 1642.860 2295.500 3105.700 1769.800 48197.00 296.2900 1985 2004.820 2541.
5、600 3866.600 2556.200 49873.00 280.5100 1986 2122.010 2763.900 4492.700 2945.600 51282.00 156.9500 1987 2199.350 3204.300 5251.600 3506.600 52783.00 212.3800 1988 2357.240 3831.000 6587.200 4510.100 54334.00 176.1800 1989 2664.900 4228.000 7278.000 5403.200 55329.00 179.4100 1990 2937.100 5017.000 7
6、717.400 5813.500 64749.00 299.5300 1991 3149.480 5228.600 9102.200 7227.000 65491.00 240.1000 1992 3483.370 5800.000 11699.50 9138.600 66152.00 265.1500 1993 4348.950 6882.100 16428.50 11323.80 66808.00 191.0400 1994 5218.100 9457.200 22372.20 14930.00 67455.00 280.1800 1995 6242.200 11993.00 28537.
7、90 17947.20 68065.00 396.1900 1996 7407.990 13844.20 33612.90 20427.50 68950.00 724.6600 1997 8651.140 14211.20 37222.70 23028.70 69820.00 682.3000 1998 9875.950 14552.40 38619.30 25173.50 70637.00 833.3000 1999 11444.08 14472.00 40557.80 27037.70 71394.00 925.4300 2000 13395.23 14628.20 44935.30 29
8、904.60 72085.00 944.9800 2001 16386.04 15411.80 48750.00 33153.00 73205.00 1218.100 2002 18903.64 16117.30 53540.70 35132.60 73740.00 1328.740 數(shù)據(jù)來源《中國統(tǒng)計年鑒——2004》 《中國統(tǒng)計年鑒——1995》 根據(jù)以上數(shù)據(jù)對模型進行回歸得 表2 Dependent Variable: Y Method: Least Squares Date: 12/11/05 Time: 13:34 Sample: 1978 2002 Included observa
9、tions: 25 Variable Coefficient Std. Error t-Statistic Prob. C -2614.437 1387.310 -1.884537 0.0749 X1 -1.213713 0.185858 -6.530340 0.0000 X2 0.437031 0.162557 2.688478 0.0145 X3 0.145897 0.198692 0.734286 0.4717 X4 0.092813 0.033189 2.796511 0.0115 X5 3.836324 0.779023 4.924535 0.0001 R-squared 0.995
10、898 Mean dependent var 5265.124 Adjusted R-squared 0.994818 S.D. dependent var 5097.514 S.E. of regression 366.9365 Akaike info criterion 14.85382 Sum squared resid 2558205. Schwarz criterion 15.14635 Log likelihood -179.6727 F-statistic 922.5526 Durbin-Watson stat 0.937935 ProbF-statistic 0.000000
11、故得出模型方程為 Y-2614.437-1.213713X10.437031X20.145897X30.092813X43.836324X5 由上表可以看出可決系數(shù)R20.995898修正后的可決系數(shù)為0.994818F統(tǒng)計量為922.5526可以看齷毓櫸匠談叨認災(zāi)荴3的T統(tǒng)計量不顯著而且X1的系數(shù)符號為負與經(jīng)濟意義不一致則各變量之間可能存在共線性 一、 對模型是否存在異方差進行判斷和處理 運用ARCH檢驗對模型進行滯后三期檢驗得 由表可知因為在給定a為0.05的條件下ObsR-squared1.726334ltX20.0537.81473所以模型不存在異方差。 ARCH Test: F-
12、statistic 0.510909 Probability 0.679822 ObsR-squared 1.726334 Probability 0.631095 Test Equation: Dependent Variable: RESID2 Method: Least Squares Date: 12/11/05 Time: 13:55 Sampleadjusted: 1981 2002 Included observations: 22 after adjusting endpoints Variable Coefficient Std. Error t-Statistic Prob
13、. C 117373.9 56618.95 2.073050 0.0528 RESID2-1 -0.242641 0.265380 -0.914315 0.3726 RESID2-2 0.036915 0.261052 0.141407 0.8891 RESID2-3 0.174483 0.290638 0.600347 0.5558 R-squared 0.078470 Mean dependent var 112254.5 Adjusted R-squared -0.075119 S.D. dependent var 121097.7 S.E. of regression 125563.7
14、 Akaike info criterion 26.48198 Sum squared resid 2.84E11 Schwarz criterion 26.68035 Log likelihood -287.3018 F-statistic 0.510909 Durbin-Watson stat 1.787966 ProbF-statistic 0.679822 二、對模型進行自相關(guān)性的檢驗與修正 在表2中我們可以看見模型的D-W統(tǒng)計量為0.937935又因為樣本容量為25解釋變量的個數(shù)為5所以在給定的顯著水平為0.05條件下根據(jù)D-W表我們可以查出上限臨界值du1.886和下限臨界值dl0
15、.953 dltdl也就是說模型存在正的一階自相關(guān)。對模型進行自相關(guān)的修正得到 Dependent Variable: Y Method: Least Squares Date: 12/11/05 Time: 14:37 Sampleadjusted: 1979 2002 Included observations: 24 after adjusting endpoints Convergence achieved after 16 iterations Variable Coefficient Std. Error t-Statistic Prob. C -1151.988 2112.532
16、 -0.545311 0.5926 X1 -1.076000 0.165946 -6.484024 0.0000 X2 0.385323 0.117262 3.286000 0.0044 X3 0.209710 0.164830 1.272277 0.2204 X4 0.056153 0.043089 1.303177 0.2099 X5 3.648192 0.771634 4.727877 0.0002 AR1 0.712070 0.276232 2.577801 0.0196 R-squared 0.997070 Mean dependent var 5437.326 Adjusted R
17、-squared 0.996035 S.D. dependent var 5132.333 S.E. of regression 323.1616 Akaike info criterion 14.63267 Sum squared resid 1775368. Schwarz criterion 14.97627 Log likelihood -168.5921 F-statistic 964.0338 Durbin-Watson stat 1.315429 ProbF-statistic 0.000000 Inverted AR Roots .71 此時D-W1.315429落在不能確定的
18、區(qū)域再次進行迭代修正得 Dependent Variable: Y Method: Least Squares Date: 12/14/05 Time: 20:57 Sampleadjusted: 1981 2002 Included observations: 22 after adjusting endpoints Convergence achieved after 13 iterations Variable Coefficient Std. Error t-Statistic Prob. C -1652.005 1008.687 -1.637777 0.1254 X1 -1.2272
19、42 0.155497 -7.892400 0.0000 X2 0.202085 0.096366 2.097059 0.0561 X3 0.543183 0.150679 3.604892 0.0032 X4 0.074132 0.023148 3.202498 0.0069 X5 3.377019 0.643299 5.249530 0.0002 AR1 0.877420 0.301336 2.911762 0.0121 AR2 -0.465979 0.388499 -1.199434 0.2518 AR3 -0.408961 0.320029 -1.277888 0.2236 R-squ
20、ared 0.998195 Mean dependent var 5826.796 Adjusted R-squared 0.997085 S.D. dependent var 5190.623 S.E. of regression 280.2630 Akaike info criterion 14.40142 Sum squared resid 1021115. Schwarz criterion 14.84776 Log likelihood -149.4156 F-statistic 898.7777 Durbin-Watson stat 2.035581 ProbF-statistic
21、 0.000000 Inverted AR Roots .64 -.76i .64.76i -.41 此時D-W為2.0355812ltdlt4-du模型的自相關(guān)性得到修正模型為 y-1652.005-1.227242x10.202085x20.543183x30.074132X4 3.377019x5 三、 對模型的多重共線性進行檢驗和修正 求相關(guān)系數(shù)距陣得出模型存在高度共線性 Correlation Matrix X1 X2 X3 X4 X5 X1 1.000000 0.984935 0.981309 0.906042 0.897899 X2 0.984935 1.000000 0.997
22、942 0.857530 0.945057 X3 0.981309 0.997942 1.000000 0.872664 0.945382 X4 0.906042 0.857530 0.872664 1.000000 0.741463 X5 0.897899 0.945057 0.945382 0.741463 1.000000 運用逐步回歸法對模型進行修正: 將各個解釋變量分別加入模型進行一元回歸: 對X1: Dependent Variable: Y Method: Least Squares Date: 12/14/05 Time: 21:09 Sample: 1978 2002 Inc
23、luded observations: 25 Variable Coefficient Std. Error t-Statistic Prob. C -653.1512 674.2466 -0.968713 0.3428 X1 0.843617 0.075976 11.10367 0.0000 R-squared 0.842780 Mean dependent var 5265.124 Adjusted R-squared 0.835944 S.D. dependent var 5097.514 S.E. of regression 2064.689 Akaike info criterion
24、 18.17997 Sum squared resid 98047615 Schwarz criterion 18.27748 Log likelihood -225.2496 F-statistic 123.2914 Durbin-Watson stat 0.208816 ProbF-statistic 0.000000 對X2: Dependent Variable: Y Method: Least Squares Date: 12/14/05 Time: 21:09 Sample: 1978 2002 Included observations: 25 Variable Coeffici
25、ent Std. Error t-Statistic Prob. C 333.8324 370.2536 0.901632 0.3766 X2 0.282231 0.015128 18.65675 0.0000 R-squared 0.938018 Mean dependent var 5265.124 Adjusted R-squared 0.935323 S.D. dependent var 5097.514 S.E. of regression 1296.384 Akaike info criterion 17.24916 Sum squared resid 38654048 Schwa
26、rz criterion 17.34667 Log likelihood -213.6145 F-statistic 348.0743 Durbin-Watson stat 0.236087 ProbF-statistic 0.000000 對X3: Dependent Variable: Y Method: Least Squares Date: 12/14/05 Time: 21:10 Sample: 1978 2002 Included observations: 25 Variable Coefficient Std. Error t-Statistic Prob. C 302.703
27、7 335.6096 0.901952 0.3764 X3 0.432024 0.020862 20.70828 0.0000 R-squared 0.949096 Mean dependent var 5265.124 Adjusted R-squared 0.946883 S.D. dependent var 5097.514 S.E. of regression 1174.830 Akaike info criterion 17.05225 Sum squared resid 31745209 Schwarz criterion 17.14976 Log likelihood -211.
28、1532 F-statistic 428.8327 Durbin-Watson stat 0.238799 ProbF-statistic 0.000000 對X4: Dependent Variable: Y Method: Least Squares Date: 12/14/05 Time: 21:10 Sample: 1978 2002 Included observations: 25 Variable Coefficient Std. Error t-Statistic Prob. C -14978.32 3308.270 -4.527538 0.0002 X4 0.344430 0
29、.055242 6.234955 0.0000 R-squared 0.628281 Mean dependent var 5265.124 Adjusted R-squared 0.612119 S.D. dependent var 5097.514 S.E. of regression 3174.736 Akaike info criterion 19.04045 Sum squared resid 2.32E08 Schwarz criterion 19.13796 Log likelihood -236.0057 F-statistic 38.87466 Durbin-Watson s
30、tat 0.134416 ProbF-statistic 0.000002 對X5: Dependent Variable: Y Method: Least Squares Date: 12/14/05 Time: 21:11 Sample: 1978 2002 Included observations: 25 Variable Coefficient Std. Error t-Statistic Prob. C -536.9009 367.2527 -1.461939 0.1573 X5 13.67779 0.663432 20.61670 0.0000 R-squared 0.94866
31、6 Mean dependent var 5265.124 Adjusted R-squared 0.946434 S.D. dependent var 5097.514 S.E. of regression 1179.781 Akaike info criterion 17.06066 Sum squared resid 32013340 Schwarz criterion 17.15817 Log likelihood -211.2583 F-statistic 425.0484 Durbin-Watson stat 1.026940 ProbF-statistic 0.000000 依據(jù)
32、可決系數(shù)最大的原則選取X3作為進入回歸模型的第一個解釋變量再依次將其余變量分別代入回歸得: 對X3、X1 Dependent Variable: Y Method: Least Squares Date: 12/14/05 Time: 21:13 Sample: 1978 2002 Included observations: 25 Variable Coefficient Std. Error t-Statistic Prob. C 1787.713 241.7849 7.393813 0.0000 X3 0.878267 0.053733 16.34510 0.0000 X1 -0.942
33、324 0.111346 -8.463043 0.0000 R-squared 0.988038 Mean dependent var 5265.124 Adjusted R-squared 0.986951 S.D. dependent var 5097.514 S.E. of regression 582.3011 Akaike info criterion 15.68402 Sum squared resid 7459640. Schwarz criterion 15.83028 Log likelihood -193.0502 F-statistic 908.6083 Durbin-W
34、atson stat 0.906175 ProbF-statistic 0.000000 對X3 X2有: Dependent Variable: Y Method: Least Squares Date: 12/14/05 Time: 21:14 Sample: 1978 2002 Included observations: 25 Variable Coefficient Std. Error t-Statistic Prob. C 310.1980 331.8091 0.934869 0.3600 X3 0.830030 0.321622 2.580759 0.0171 X2 -0.26
35、2077 0.211345 -1.240044 0.2280 R-squared 0.952422 Mean dependent var 5265.124 Adjusted R-squared 0.948096 S.D. dependent var 5097.514 S.E. of regression 1161.334 Akaike info criterion 17.06469 Sum squared resid 29671306 Schwarz criterion 17.21096 Log likelihood -210.3087 F-statistic 220.1980 Durbin-
36、Watson stat 0.288980 ProbF-statistic 0.000000 對X3 X4有: Dependent Variable: Y Method: Least Squares Date: 12/14/05 Time: 21:14 Sample: 1978 2002 Included observations: 25 Variable Coefficient Std. Error t-Statistic Prob. C 5391.098 1796.194 3.001400 0.0066 X3 0.525376 0.037257 14.10155 0.0000 X4 -0.1
37、04820 0.036507 -2.871246 0.0089 R-squared 0.962972 Mean dependent var 5265.124 Adjusted R-squared 0.959606 S.D. dependent var 5097.514 S.E. of regression 1024.517 Akaike info criterion 16.81400 Sum squared resid 23091963 Schwarz criterion 16.96026 Log likelihood -207.1750 F-statistic 286.0707 Durbin
38、-Watson stat 0.352644 ProbF-statistic 0.000000 對X3 X5有: Dependent Variable: Y Method: Least Squares Date: 12/14/05 Time: 21:15 Sample: 1978 2002 Included observations: 25 Variable Coefficient Std. Error t-Statistic Prob. C -266.5145 265.0751 -1.005430 0.3256 X3 0.222948 0.045379 4.913017 0.0001 X5 7
39、.003328 1.437013 4.873531 0.0001 R-squared 0.975522 Mean dependent var 5265.124 Adjusted R-squared 0.973297 S.D. dependent var 5097.514 S.E. of regression 832.9854 Akaike info criterion 16.40008 Sum squared resid 15265021 Schwarz criterion 16.54634 Log likelihood -202.0009 F-statistic 438.3899 Durbi
40、n-Watson stat 0.710798 ProbF-statistic 0.000000 在滿足經(jīng)濟意義和可決系數(shù)的條件下選取X5作為進入模型的第二個解釋變量再次進行回歸則: 對X3 X5 X1有: Dependent Variable: Y Method: Least Squares Date: 12/14/05 Time: 21:19 Sample: 1978 2002 Included observations: 25 Variable Coefficient Std. Error t-Statistic Prob. C 1158.891 219.0677 5.290105 0.0
41、000 X3 0.668525 0.059040 11.32327 0.0000 X5 3.827440 0.817330 4.682856 0.0001 X1 -0.740704 0.090594 -8.176061 0.0000 R-squared 0.994149 Mean dependent var 5265.124 Adjusted R-squared 0.993313 S.D. dependent var 5097.514 S.E. of regression 416.8529 Akaike info criterion 15.04899 Sum squared resid 364
42、9093. Schwarz criterion 15.24401 Log likelihood -184.1124 F-statistic 1189.303 Durbin-Watson stat 0.966444 ProbF-statistic 0.000000 對X3 X5 X2有: Dependent Variable: Y Method: Least Squares Date: 12/14/05 Time: 21:20 Sample: 1978 2002 Included observations: 25 Variable Coefficient Std. Error t-Statist
43、ic Prob. C -276.5219 242.2584 -1.141434 0.2665 X3 0.703832 0.212045 3.319260 0.0033 X5 7.239493 1.317076 5.496638 0.0000 X2 -0.321294 0.138938 -2.312490 0.0310 R-squared 0.980490 Mean dependent var 5265.124 Adjusted R-squared 0.977703 S.D. dependent var 5097.514 S.E. of regression 761.1637 Akaike in
44、fo criterion 16.25322 Sum squared resid 12166775 Schwarz criterion 16.44824 Log likelihood -199.1653 F-statistic 351.7985 Durbin-Watson stat 1.048322 ProbF-statistic 0.000000 對X3 X5 X4有: Dependent Variable: Y Method: Least Squares Date: 12/14/05 Time: 21:20 Sample: 1978 2002 Included observations: 2
45、5 Variable Coefficient Std. Error t-Statistic Prob. C 1752.237 1771.385 0.989191 0.3338 X3 0.288730 0.072713 3.970791 0.0007 X5 5.989815 1.675741 3.574429 0.0018 X4 -0.039889 0.034613 -1.152431 0.2621 R-squared 0.976978 Mean dependent var 5265.124 Adjusted R-squared 0.973690 S.D. dependent var 5097.
46、514 S.E. of regression 826.8426 Akaike info criterion 16.41875 Sum squared resid 14357043 Schwarz criterion 16.61377 Log likelihood -201.2344 F-statistic 297.0613 Durbin-Watson stat 0.603278 ProbF-statistic 0.000000 可見加入其余任何一個變量都會導(dǎo)致系數(shù)符號與經(jīng)濟意義不符故最終修正后的回歸模型為: Dependent Variable: Y Method: Least Squares
47、 Date: 12/14/05 Time: 21:23 Sample: 1978 2002 Included observations: 25 Variable Coefficient Std. Error t-Statistic Prob. C -266.5145 265.0751 -1.005430 0.3256 X3 0.222948 0.045379 4.913017 0.0001 X5 7.003328 1.437013 4.873531 0.0001 R-squared 0.975522 Mean dependent var 5265.124 Adjusted R-squared 0.973297 S.D. dependent var 5097.514 S.
對我國財政收入影響因素的回歸分析 計量經(jīng)濟學(xué)論文
