經(jīng)濟(jì)計(jì)量學(xué)實(shí)習(xí)報(bào)告影響我國(guó)農(nóng)村居民消費(fèi)水平的主要因素分析

1、 經(jīng)濟(jì)計(jì)量學(xué)實(shí)習(xí)報(bào)告影響我國(guó)農(nóng)村居民消費(fèi)水平的主要因素分析 【摘要】：本文主要通過對(duì)農(nóng)村居民消費(fèi)水平的變動(dòng)進(jìn)行多因素分析，建立以農(nóng)村居民消費(fèi)水平為應(yīng)變量，以農(nóng)村人口自然增長(zhǎng)率、農(nóng)村居民人均可支配收入、商品零售價(jià)格指數(shù)以及農(nóng)業(yè)生產(chǎn)資料價(jià)格指數(shù)為自變量的多元線性回歸模型，并利用模型對(duì)農(nóng)村居民消費(fèi)水平這一社會(huì)現(xiàn)象進(jìn)行數(shù)量化分析，揭示中國(guó)農(nóng)村消費(fèi)水平的現(xiàn)狀及問題，并對(duì)如何提高農(nóng)村居民消費(fèi)水平提出一些可行性的建議?！娟P(guān)鍵詞】:農(nóng)村居民消費(fèi)水平、農(nóng)村人口自然增長(zhǎng)率、農(nóng)村居民人均可支配收入、商品零售價(jià)格指數(shù)、農(nóng)業(yè)生產(chǎn)資料價(jià)格指數(shù)、建議前言: 當(dāng)前全球面臨60 年來最嚴(yán)重的金融危機(jī)之后的經(jīng)濟(jì)復(fù)蘇期，而中國(guó)亦深

2、受當(dāng)前經(jīng)濟(jì)時(shí)勢(shì)影響，外貿(mào)出口難度加大。我國(guó)地域遼闊，經(jīng)濟(jì)發(fā)展不平衡，人民生活由溫飽向小康過渡，無論是市場(chǎng)容量還是未來發(fā)展，擴(kuò)大內(nèi)需的潛力都十分巨大。此外，當(dāng)前工業(yè)化，城市化，現(xiàn)代化進(jìn)程加快，經(jīng)濟(jì)結(jié)構(gòu)調(diào)整升級(jí)，國(guó)內(nèi)市場(chǎng)的需求進(jìn)一步擴(kuò)大。所以，對(duì)我國(guó)這樣一個(gè)發(fā)展中大國(guó)來說，拉動(dòng)經(jīng)濟(jì)增長(zhǎng)的最主要力量仍然是國(guó)內(nèi)需求， 而擴(kuò)大國(guó)內(nèi)需求的一個(gè)重要舉措是刺激國(guó)內(nèi)消費(fèi)。而農(nóng)民作為中國(guó)廣大的消費(fèi)群體，其消費(fèi)水平和消費(fèi)需求的變化直接關(guān)系到內(nèi)需的政策的效果。目前，農(nóng)民的經(jīng)濟(jì)狀況仍然保持在“溫飽有余、小康不足”的狀態(tài)?！霸S多農(nóng)民消費(fèi)仍然不足，這已經(jīng)影響到整個(gè)國(guó)民經(jīng)濟(jì)的健康發(fā)展。因此研究中國(guó)農(nóng)村居民消費(fèi)水平，對(duì)于我國(guó)制

3、定、完善經(jīng)濟(jì)政策，改善消費(fèi)結(jié)構(gòu)，促進(jìn)消費(fèi)水平，提高農(nóng)民消費(fèi)質(zhì)量有重要的意義。一、數(shù)據(jù)整理以及模型預(yù)測(cè)影響我國(guó)農(nóng)村居民消費(fèi)水平的主要因素分析年份農(nóng)村居民消費(fèi)水平y(tǒng)（元）農(nóng)村人口自然增長(zhǎng)率x1農(nóng)村居民人均可支配收入x2（元)商品零售價(jià)格指數(shù)x3農(nóng)業(yè)生產(chǎn)資料價(jià)格指數(shù)x4198534914.26397.6108.8104.8198637815.57423.8106101.1198742116.61462.6107.3107198850915.73544.9118.5116.2198954915.04601.5117.8118.9199056014.39686.3102.1105.5199160212.

4、98708.6102.9102.9199268811.6784105.4103.7199380511.45921.6113.2114.11994103811.211221121.7121.61995131310.551577.7114.8127.41996162610.421926.1106.1108.41997172210.062090.1100.899.5199817309.14216297.494.5199917668.182210.39795.8200018607.582253.498.599.1200119696.952366.499.299.1200220626.452475.69

5、8.7100.5200321036.012622.299.9101.4200423015.872936.4102.8110.6200525605.893254.9100.8108.3200628475.283587101101.5200732655.174140.4103.8107.7200837565.084760.6105.9120.3200942505.055153101.497.5 數(shù)據(jù)來源：2009年中國(guó)統(tǒng)計(jì)年鑒根據(jù)上面的數(shù)據(jù)我們初步預(yù)測(cè)模型：y=b0+b1*x1+b2*x2+b3*x3+b4*x4+u其中：y農(nóng)村居民消費(fèi)水平x1農(nóng)村人口自然增長(zhǎng)率x2農(nóng)村居民人均可支配收入x3商品零

6、售價(jià)格指數(shù)x4農(nóng)業(yè)生產(chǎn)資料價(jià)格指數(shù)u隨機(jī)誤差項(xiàng)二、模型設(shè)定回歸模型參數(shù)估計(jì)根據(jù)數(shù)據(jù)用eviews軟件對(duì)模型進(jìn)行ols估計(jì)，得樣本回歸方程。結(jié)果如下：dependent variable: ymethod: least squaresdate: 01/15/11 time: 20:48sample: 1985 2009included observations: 25variablecoefficientstd. errort-statisticprob. c-207.4927163.4764-1.2692520.2189x1-0.7415855.960389-0.1244190.9022x20

7、.7986720.01516052.684330.0000x37.9035243.1062042.5444310.0193x4-5.4543712.110245-2.5847090.0177r-squared0.998772 mean dependent var1641.160adjusted r-squared0.998526 s.d. dependent var1095.098s.e. of regression42.04608 akaike info criterion10.49227sum squared resid35357.46 schwarz criterion10.73604l

8、og likelihood-126.1533 f-statistic4065.108durbin-watson stat1.025801 prob(f-statistic)0.000000經(jīng)過上述的初步回歸分析，表明了最小的二乘估計(jì)的性質(zhì)，證明了最小二乘法準(zhǔn)則的合理性，但仍然不能完全保證現(xiàn)行回歸分析的價(jià)值。原因是，模型本身未必一定滿足要求，也就是模型的各個(gè)假設(shè)并不一定成立。最終的結(jié)果為：i = -207.4927-0.741585*x1+0.798672*x2+ 7.903524*x3-5.454371*x4t= (-1.269252) (-0.124419) (52.68433) (2.54

9、4431) (-2.584709) r2=0.998772 r2=0.998526 dw=1.025801 f=4065.108模型檢驗(yàn)：經(jīng)濟(jì)意義檢驗(yàn)：從得出的模型看，x1和x4的參數(shù)符號(hào)沒通過經(jīng)濟(jì)意義檢驗(yàn)。r2檢驗(yàn)：經(jīng)計(jì)算此模型的可決系數(shù)r2=0.998772，校正的可決系數(shù)r2=0.998526，表明模型擬合度高。t檢驗(yàn):再從五個(gè)參數(shù)的t檢驗(yàn)值看，五個(gè)參數(shù)的t值分別為：t0=-1.269252, t1=-0.124419, t2=52.68433, t3=2.544431, t4=-2.584709 ，在5%顯著性水平下自由度為n-k=25-5=20的t分布臨界值為2.086，因此可知有部

10、分t值是不顯著的。f檢驗(yàn)：模型的f值為：f=4065.108，而5%顯著性水平下自由度分別為k-1=4和n-k=20的f分布臨界值遠(yuǎn)小于模型的f值，說明模型在總體上是高度顯著的。下面進(jìn)行相關(guān)檢驗(yàn)說明模型中可能存在多重共線性等問題，進(jìn)而對(duì)模型進(jìn)行修正。三、模型的檢驗(yàn)和修正1.多重共線性檢驗(yàn):yx1x2x3x4y1-0.9081173817440.999151350318-0.45713591375-0.160979829436x1-0.9081173817441-0.9113872994820.5581729288970.231994327899x20.999151350318-0.911387

11、2994821-0.463297625812-0.156835642896x3-0.457135913750.558172928897-0.46329762581210.837520186067x4-0.1609798294360.231994327899-0.1568356428960.8375201860671由上表可知，x1與x2相關(guān)系數(shù)高達(dá)0.9114，x4與x3相關(guān)系數(shù)高達(dá)0.8375，結(jié)合經(jīng)濟(jì)意義和統(tǒng)計(jì)檢驗(yàn)選出擬合效果最好的一元線性回歸方程。多重共線修正處理：(1)采用逐步回歸: 運(yùn)用ols方法求y對(duì)各個(gè)解釋變量的回歸。結(jié)合經(jīng)濟(jì)意義和統(tǒng)計(jì)檢驗(yàn)選出擬合效果最好的一元線性回歸方程。ev

12、iews過程如下y對(duì)x1回歸：dependent variable: ymethod: least squaresdate: 01/15/11 time: 20:50sample: 1985 2009included observations: 25variablecoefficientstd. errort-statisticprob. c4168.325260.401016.007330.0000x1-256.284024.63966-10.401280.0000r-squared0.824677 mean dependent var1641.160adjusted r-squared0.

13、817054 s.d. dependent var1095.098s.e. of regression468.3967 akaike info criterion15.21313sum squared resid5046096. schwarz criterion15.31064log likelihood-188.1641 f-statistic108.1866durbin-watson stat0.281126 prob(f-statistic)0.000000y對(duì)x2回歸：dependent variable: ymethod: least squaresdate: 01/15/11 t

14、ime: 20:50sample: 1985 2009included observations: 25variablecoefficientstd. errort-statisticprob. c57.4474316.439243.4945300.0020x20.7876350.006770116.33440.0000r-squared0.998303 mean dependent var1641.160adjusted r-squared0.998230 s.d. dependent var1095.098s.e. of regression46.07673 akaike info cri

15、terion10.57511sum squared resid48830.50 schwarz criterion10.67262log likelihood-130.1889 f-statistic13533.69durbin-watson stat1.178851 prob(f-statistic)0.000000y對(duì)x3回歸：dependent variable: ymethod: least squaresdate: 01/15/11 time: 20:50sample: 1985 2009included observations: 25variablecoefficientstd.

16、 errort-statisticprob. c9234.5533086.9272.9915030.0065x3-72.1311729.26236-2.4649820.0216r-squared0.208973 mean dependent var1641.160adjusted r-squared0.174581 s.d. dependent var1095.098s.e. of regression994.9247 akaike info criterion16.71983sum squared resid22767128 schwarz criterion16.81734log like

17、lihood-206.9979 f-statistic6.076134durbin-watson stat0.212411 prob(f-statistic)0.021595y對(duì)x4回歸：dependent variable: ymethod: least squaresdate: 01/15/11 time: 20:51sample: 1985 2009included observations: 25variablecoefficientstd. errort-statisticprob. c3798.6572766.9471.3728700.1830x4-20.2209825.85028

18、-0.7822340.4421r-squared0.025915 mean dependent var1641.160adjusted r-squared-0.016437 s.d. dependent var1095.098s.e. of regression1104.061 akaike info criterion16.92800sum squared resid28035877 schwarz criterion17.02551log likelihood-209.6000 f-statistic0.611890durbin-watson stat0.062338 prob(f-sta

19、tistic)0.442056從上述四個(gè)表格分析可以得出：y對(duì)x2的線性關(guān)系強(qiáng)，擬合程度最優(yōu)，則有回歸方程：y=57.44743+0.787635*x2（2）逐步回歸，將其余解釋變量逐一代入上式引入x1：dependent variable: ymethod: least squaresdate: 01/16/11 time: 01:33sample: 1985 2009included observations: 25variablecoefficientstd. errort-statisticprob.c-4.86309690.75772-0.0535830.9578x14.159691

20、5.9562290.6983770.4923x20.7982240.01663747.977700.0000r-squared0.998340mean dependent var1641.160adjusted r-squared0.998189s.d. dependent var1095.098s.e. of regression46.59859akaike info criterion10.63318sum squared resid47771.43schwarz criterion10.77945log likelihood-129.9148hannan-quinn criter.10.

21、67375f-statistic6616.374durbin-watson stat1.185921prob(f-statistic)0.000000引進(jìn)x3:dependent variable: ymethod: least squaresdate: 01/15/11 time: 20:53sample: 1985 2009included observations: 25variablecoefficientstd. errort-statisticprob. c-69.95493170.5280-0.4102250.6856x20.7903170.007713102.46070.000

22、0x31.1589831.5439330.7506690.4608r-squared0.998346 mean dependent var1641.160adjusted r-squared0.998195 s.d. dependent var1095.098s.e. of regression46.52028 akaike info criterion10.62982sum squared resid47611.00 schwarz criterion10.77609log likelihood-129.8728 f-statistic6638.706durbin-watson stat1.

23、209349 prob(f-statistic)0.000000引進(jìn)x4dependent variable: ymethod: least squaresdate: 01/16/11 time: 01:34sample: 1985 2009included observations: 25variablecoefficientstd. errort-statisticprob.c117.3086121.85830.9626640.3462x20.7870920.006970112.91780.0000x4-0.5508271.110709-0.4959230.6249r-squared0.9

24、98322mean dependent var1641.160adjusted r-squared0.998170s.d. dependent var1095.098s.e. of regression46.85115akaike info criterion10.64399sum squared resid48290.66schwarz criterion10.79026log likelihood-130.0499hannan-quinn criter.10.68456f-statistic6545.116durbin-watson stat1.113382prob(f-statistic

25、)0.000000經(jīng)上面的分析，再次依據(jù)調(diào)整后的可決系數(shù)最大原則，選取調(diào)整后可決系數(shù)最大所對(duì)應(yīng)的解釋變量作為新進(jìn)入模型的候選變量，將這個(gè)候選變量的調(diào)整后可決系數(shù)與上一步中進(jìn)入模型解釋變量的調(diào)整后可決系數(shù)加以比較，若是大于上一步的調(diào)整后可決系數(shù)，則將候選變量加入模型，若是小于，則將停止逐步回歸。經(jīng)查x3的調(diào)整后可決系數(shù)最大，故x3作為第二個(gè)解釋變量進(jìn)入回歸模型。（3）繼續(xù)逐步回歸加入x1dependent variable: ymethod: least squaresdate: 01/15/11 time: 22:43sample: 1985 2009included observations

26、: 25variablecoefficientstd. errort-statisticprob. c-82.54114176.0244-0.4689190.6440x12.8987446.5281120.4440400.6616x20.7970470.01707346.685460.0000x30.8784681.6950210.5182640.6097r-squared0.998361 mean dependent var1641.160adjusted r-squared0.998127 s.d. dependent var1095.098s.e. of regression47.393

27、05 akaike info criterion10.70048sum squared resid47168.13 schwarz criterion10.89550log likelihood-129.7559 f-statistic4264.362durbin-watson stat1.202271 prob(f-statistic)0.000000加入x4dependent variable: ymethod: least squaresdate: 01/16/11 time: 01:35sample: 1985 2009included observations: 25variable

28、coefficientstd. errort-statisticprob.c-209.1116159.0920-1.3144070.2029x20.8002790.007746103.31550.0000x37.7558612.8024382.7675410.0115x4-5.3923312.001844-2.6936820.0136r-squared0.998771mean dependent var1641.160adjusted r-squared0.998595s.d. dependent var1095.098s.e. of regression41.04865akaike info

29、 criterion10.41304sum squared resid35384.83schwarz criterion10.60806log likelihood-126.1630hannan-quinn criter.10.46713f-statistic5686.745durbin-watson stat1.013575prob(f-statistic)0.000000根據(jù)上面的表格可知，由于，此次調(diào)整后可決系數(shù)最大的是x4，但與上一步的調(diào)整后可決系數(shù)相比要小，故可以認(rèn)為逐步回歸終止。由于在這里加入x4這個(gè)變量對(duì)于r2影響幾乎不可計(jì)算，因此在這里不將x4放入模型。故修正后的模型是： y=

30、 -207.4927 +0.798672*x2+ 7.903524*x3y農(nóng)村居民消費(fèi)水平x2農(nóng)村居民人均可支配收入x3商品零售價(jià)格指數(shù)x4農(nóng)業(yè)生產(chǎn)資料價(jià)格指數(shù)2.異方差檢驗(yàn)異方差分析：g-q檢驗(yàn)：去除中間3個(gè)數(shù)據(jù)，剩下22個(gè)數(shù)據(jù)，此時(shí)自由度為11-2-1=8，查表的出臨界值（9，9）=3.18對(duì)x2進(jìn)行排序可得：子樣一：dependent variable: ymethod: least squaresdate: 01/15/11 time: 21:01sample: 1 11included observations: 11variablecoefficientstd. errort-st

31、atisticprob. c-192.903276.94086-2.5071620.0365x20.7942360.01350158.828110.0000x32.2308310.7268763.0690680.0154r-squared0.998177 mean dependent var655.6364adjusted r-squared0.997722 s.d. dependent var295.4225s.e. of regression14.10154 akaike info criterion8.357447sum squared resid1590.829 schwarz cri

32、terion8.465963log likelihood-42.96596 f-statistic2190.440durbin-watson stat1.510644 prob(f-statistic)0.000000子樣二：dependent variable: ymethod: least squaresdate: 01/15/11 time: 21:02sample: 15 25included observations: 11variablecoefficientstd. errort-statisticprob. c2656.866899.44402.9538990.0183x20.

33、8482010.02369435.798680.0000x3-27.789339.499651-2.9253000.0191r-squared0.997180 mean dependent var2612.636adjusted r-squared0.996475 s.d. dependent var826.3828s.e. of regression49.06500 akaike info criterion10.85117sum squared resid19258.99 schwarz criterion10.95969log likelihood-56.68143 f-statisti

34、c1414.368durbin-watson stat1.081785 prob(f-statistic)0.000000從上兩個(gè)表格求得：f1=19258.99/1590.829=12.10對(duì)x3進(jìn)行排序：子樣一：dependent variable: ymethod: least squaresdate: 01/15/11 time: 21:04sample: 1 11included observations: 11variablecoefficientstd. errort-statisticprob. c65.703321048.5600.0626610.9516x20.814236

35、0.01595951.020520.0000x3-0.72351610.59494-0.0682890.9472r-squared0.997069 mean dependent var2129.909adjusted r-squared0.996336 s.d. dependent var906.7206s.e. of regression54.88365 akaike info criterion11.07531sum squared resid24097.72 schwarz criterion11.18383log likelihood-57.91420 f-statistic1360.

36、681durbin-watson stat1.807882 prob(f-statistic)0.000000子樣二：dependent variable: ymethod: least squaresdate: 01/15/11 time: 21:06sample: 15 25included observations: 11variablecoefficientstd. errort-statisticprob. c-69.27762161.1354-0.4299340.6786x20.7798430.006715116.13150.0000x31.2823331.4240730.9004

37、690.3942r-squared0.999442 mean dependent var1039.273adjusted r-squared0.999303 s.d. dependent var989.7291s.e. of regression26.13108 akaike info criterion9.591129sum squared resid5462.668 schwarz criterion9.699646log likelihood-49.75121 f-statistic7168.785durbin-watson stat2.490933 prob(f-statistic)0

38、.000000根據(jù)上兩表求得：f2=24097.72/5462.668=4.41因?yàn)閒1f2（9，9）=3.18 ，所以模型存在異方差。需要對(duì)其進(jìn)行修正。異方差修正：dependent variable: ymethod: least squaresdate: 01/15/11 time: 21:09sample: 1985 2009included observations: 25weighting series: 1/abs(el)variablecoefficientstd. errort-statisticprob. c-100.421318.87623-5.3199890.0000x

39、20.7875570.002494315.81420.0000x31.4632360.1808788.0896130.0000weighted statisticsr-squared0.999957 mean dependent var893.7122adjusted r-squared0.999953 s.d. dependent var1198.539s.e. of regression8.227519 akaike info criterion7.165013sum squared resid1489.225 schwarz criterion7.311278log likelihood

40、-86.56266 f-statistic93346.88durbin-watson stat1.292238 prob(f-statistic)0.000000unweighted statisticsr-squared0.998310 mean dependent var1641.160adjusted r-squared0.998156 s.d. dependent var1095.098s.e. of regression47.02431 sum squared resid48648.29durbin-watson stat1.198471所以，修正后的模型為：y =-100.4213

41、+0.787557*x2+1.463236*x3t=（-5.319989） （315.8142） （8.089613） r2=0.999957 f=93346.883.自相關(guān)性檢驗(yàn)：dependent variable: ymethod: least squaresdate: 01/15/11 time: 20:53sample: 1985 2009included observations: 25variablecoefficientstd. errort-statisticprob. c-69.95493170.5280-0.4102250.6856x20.7903170.00771310

42、2.46070.0000x31.1589831.5439330.7506690.4608r-squared0.998346 mean dependent var1641.160adjusted r-squared0.998195 s.d. dependent var1095.098s.e. of regression46.52028 akaike info criterion10.62982sum squared resid47611.00 schwarz criterion10.77609log likelihood-129.8728 f-statistic6638.706durbin-watson stat1.209349 prob(f-statistic)0.000000查表可得其臨界值為：dl=1.21 ， du=1.55 ，此時(shí)，d.w=1.209349dl=1.21 由此可見，模型存在正自相關(guān)。4-du=2.45自相關(guān)修正：dependent variable: ymethod: least squaresdate: 01/16/11 time: 02:5