panelinių duomenų modeliavimas su eviews
DESCRIPTION
Panelinių duomenų modeliavimas su Eviews. 2013-12-11. Pasirinkti nagrinėjamą reiškinį. Surinkti reikiamus duomenis. Perkelti juos į Eviews programą. Atlikti duomenų grafinę ir statistinę analizę. - PowerPoint PPT PresentationTRANSCRIPT
Panelinių duomenų modeliavimas su Eviews
2013-12-11
1. Pasirinkti nagrinėjamą reiškinį. Surinkti reikiamus duomenis. Perkelti juos į Eviews programą.
2. Atlikti duomenų grafinę ir statistinę analizę. 3. Apskaičiuoti pastovios konstantos, fiksuotų ir atsitiktinių efektų
panelinius modelius4. Įsitikinti, kuris modelio sudarymo metodas tinkamiausias ir
kodėl?1. F-testas2. Hausman testas
5. Patikrinti modelio specifikaciją ir determinuotumą. 1. Omitted/Rdundnt variables test
2. Determinuotumo rodikliai
6. Modelio bendra išraiška ir koeficientų interpretacija
Apskaičiuoti pastovios konstantos, fiksuotų ir atsitiktinių efektų panelinius modelius
• Pastovios konstantos
12
Apskaičiuoti pastovios konstantos, fiksuotų ir atsitiktinių efektų panelinius modelius
• Pastovios konstantos
Priklausomas kint.
Įtaką darantys veiskn
Pastovios konst.
Apskaičiuoti pastovios konstantos, fiksuotų ir atsitiktinių efektų panelinius modelius
• Pastovios konstantosDependent Variable: UZ_?Method: Pooled Least SquaresDate: 12/10/13 Time: 22:31Sample (adjusted): 2008Q3 2013Q1Included observations: 19 after adjustmentsCross-sections included: 15Total pool (balanced) observations: 285
Variable Coefficient Std. Error t-Statistic Prob.
C 61165.29 242358.4 0.252375 0.8009AUGIMAS_? -77.19613 124.9949 -0.617594 0.5373
MMA(-2) -7.283652 298.8090 -0.024376 0.9806KRIZE -3963.055 6654.056 -0.595585 0.5519
R-squared 0.002562 Mean dependent var 45533.21Adjusted R-squared -0.008087 S.D. dependent var 54336.17S.E. of regression 54555.43 Akaike info criterion 24.66576Sum squared resid 8.36E+11 Schwarz criterion 24.71702Log likelihood -3510.870 Hannan-Quinn criter. 24.68631F-statistic 0.240584 Durbin-Watson stat 0.007690Prob(F-statistic) 0.867992
Fiksuotų efektų (FE) panelinius modelius
Fiksuoti skerspjūvio efektai
Fiksuotų efektų (FE) panelinius modelius
Fiksuoti skerspjūvio efektai
Dependent Variable: UZ_?Method: Pooled Least SquaresDate: 12/10/13 Time: 22:41Sample (adjusted): 2008Q3 2013Q1Included observations: 19 after adjustmentsCross-sections included: 15Total pool (balanced) observations: 285
Variable Coefficient Std. Error t-Statistic Prob.
C 35271.02 27607.27 1.277599 0.2025AUGIMAS_? 4.471226 14.99697 0.298142 0.7658
MMA(-2) 14.38877 34.01001 0.423075 0.6726KRIZE -3683.400 757.0060 -4.865748 0.0000
Fixed Effects (Cross)A--C -27028.13B--C 102027.3C--C 99689.09D--C -39714.94E--C -42130.05G--C 19646.74H--C 107290.2I--C 524.2334J--C -24729.77K--C -27611.48M--C -31074.07O--C -36585.86P--C -22854.95Q--C -41589.67R--C -35858.60
Effects Specification
Cross-section fixed (dummy variables)
R-squared 0.987739 Mean dependent var 45533.21Adjusted R-squared 0.986959 S.D. dependent var 54336.17S.E. of regression 6205.111 Akaike info criterion 20.36521Sum squared resid 1.03E+10 Schwarz criterion 20.59589Log likelihood -2884.042 Hannan-Quinn criter. 20.45768F-statistic 1265.292 Durbin-Watson stat 0.293817Prob(F-statistic) 0.000000
F testas
F testas
F testas Fiksuoti efektai yra pertekliniai
Redundant Fixed Effects TestsPool: DU_SAKOSTest cross-section fixed effects
Effects Test Statistic d.f. Prob.
Cross-section F 1532.440475 (14,267) 0.0000Cross-section Chi-square 1253.656241 14 0.0000
Cross-section fixed effects test equation:Dependent Variable: UZ_?Method: Panel Least SquaresDate: 12/10/13 Time: 22:52Sample (adjusted): 2008Q3 2013Q1Included observations: 19 after adjustmentsCross-sections included: 15Total pool (balanced) observations: 285
Variable Coefficient Std. Error t-Statistic Prob.
C 61165.29 242358.4 0.252375 0.8009AUGIMAS_? -77.19613 124.9949 -0.617594 0.5373
MMA(-2) -7.283652 298.8090 -0.024376 0.9806KRIZE -3963.055 6654.056 -0.595585 0.5519
R-squared 0.002562 Mean dependent var 45533.21Adjusted R-squared -0.008087 S.D. dependent var 54336.17S.E. of regression 54555.43 Akaike info criterion 24.66576Sum squared resid 8.36E+11 Schwarz criterion 24.71702Log likelihood -3510.870 Hannan-Quinn criter. 24.68631F-statistic 0.240584 Durbin-Watson stat 0.007690Prob(F-statistic) 0.867992
H0: atmetama
Atsitiktinių efektų (FE)modelius
Atsitiktiniai skerspjūvio efektai
Atsitiktinių efektų modelisDependent Variable: UZ_?Method: Pooled EGLS (Cross-section random effects)Date: 12/10/13 Time: 23:05Sample (adjusted): 2008Q3 2013Q1Included observations: 19 after adjustmentsCross-sections included: 15Total pool (balanced) observations: 285Swamy and Arora estimator of component variances
Variable Coefficient Std. Error t-Statistic Prob.
C 35288.72 31335.94 1.126142 0.2611AUGIMAS_? 4.415402 14.99645 0.294430 0.7686
MMA(-2) 14.37396 34.00999 0.422639 0.6729KRIZE -3683.591 757.0059 -4.866001 0.0000
Random Effects (Cross)A--C -27009.82B--C 101964.5C--C 99627.73D--C -39690.67E--C -42104.29G--C 19634.67H--C 107224.1I--C 523.7777J--C -24714.71K--C -27594.65M--C -31055.09O--C -36563.52P--C -22841.07Q--C -41564.27R--C -35836.70
Effects SpecificationS.D. Rho
Cross-section random 57417.07 0.9885Idiosyncratic random 6205.111 0.0115
Weighted Statistics
R-squared 0.086088 Mean dependent var 1128.565Adjusted R-squared 0.076331 S.D. dependent var 6447.217S.E. of regression 6196.273 Sum squared resid 1.08E+10F-statistic 8.823117 Durbin-Watson stat 0.280018Prob(F-statistic) 0.000013
Unweighted Statistics
R-squared 0.001049 Mean dependent var 45533.21Sum squared resid 8.38E+11 Durbin-Watson stat 0.003607
FE ir AE palyginimasHausmas testas Atsitiktiniai efektai
Hausman testas
Hausman testasCorrelated Random Effects - Hausman TestPool: DU_SAKOSTest cross-section random effects
Test Summary Chi-Sq. Statistic Chi-Sq. d.f. Prob.
Cross-section random 0.000000 3 1.0000
* Cross-section test variance is invalid. Hausman statistic set to zero.
Cross-section random effects test comparisons:
Variable Fixed Random Var(Diff.) Prob.
AUGIMAS_? 4.471226 4.415402 0.015578 0.6547MMA(-2) 14.388771 14.373957 0.001097 0.6547KRIZE -3683.400... -3683.5913... 0.182666 0.6547
Cross-section random effects test equation:Dependent Variable: UZ_?Method: Panel Least SquaresDate: 12/10/13 Time: 23:11Sample (adjusted): 2008Q3 2013Q1Included observations: 19 after adjustmentsCross-sections included: 15Total pool (balanced) observations: 285
Variable Coefficient Std. Error t-Statistic Prob.
C 35271.02 27607.27 1.277599 0.2025AUGIMAS_? 4.471226 14.99697 0.298142 0.7658
MMA(-2) 14.38877 34.01001 0.423075 0.6726KRIZE -3683.400 757.0060 -4.865748 0.0000
Effects Specification
Cross-section fixed (dummy variables)
R-squared 0.987739 Mean dependent var 45533.21Adjusted R-squared 0.986959 S.D. dependent var 54336.17S.E. of regression 6205.111 Akaike info criterion 20.36521Sum squared resid 1.03E+10 Schwarz criterion 20.59589Log likelihood -2884.042 Hannan-Quinn criter. 20.45768F-statistic 1265.292 Durbin-Watson stat 0.293817Prob(F-statistic) 0.000000
Paieška geresnio modelio (be autokoreliacijos )
Paieška geresnio modelio
Paieška geresnio modelio Dependent Variable: DLOG(UZ_?)Method: Pooled Least SquaresDate: 12/10/13 Time: 23:37Sample (adjusted): 2008Q4 2013Q1Included observations: 18 after adjustmentsCross-sections included: 15Total pool (balanced) observations: 270
Variable Coefficient Std. Error t-Statistic Prob.
C -0.165381 0.019173 -8.625649 0.0000AUGIMAS_? 0.000828 0.000122 6.779932 0.0000
DLOG(MMA(-2)) 0.420724 0.271569 1.549239 0.1226@TREND 0.013352 0.003245 4.114071 0.0001
@TREND^2 -0.000478 0.000144 -3.307822 0.0011Fixed Effects (Cross)
A--C -0.026977B--C -0.004479C--C -0.004290D--C -0.014311E--C 0.006178G--C -0.027630H--C -0.002373I--C 0.011694J--C -0.004159K--C 0.008147M--C -0.009978O--C 0.006032P--C 0.012613Q--C 0.024438R--C 0.025096
Effects Specification
Cross-section fixed (dummy variables)
R-squared 0.302266 Mean dependent var -0.001930Adjusted R-squared 0.252229 S.D. dependent var 0.056786S.E. of regression 0.049105 Akaike info criterion -3.121958Sum squared resid 0.605226 Schwarz criterion -2.868735Log likelihood 440.4643 Hannan-Quinn criter. -3.020275F-statistic 6.040882 Durbin-Watson stat 2.109491Prob(F-statistic) 0.000000
Paklaidų diagrama
-.12
-.08
-.04
.00
.04
.08
IV I II III IV I II III IV I II III IV I II III IV I
2009 2010 2011 2012 2013
A Residuals
-.06
-.04
-.02
.00
.02
.04
.06
IV I II III IV I II III IV I II III IV I II III IV I
2009 2010 2011 2012 2013
B Residuals
-.06
-.04
-.02
.00
.02
.04
.06
IV I II III IV I II III IV I II III IV I II III IV I
2009 2010 2011 2012 2013
C Residuals
-.3
-.2
-.1
.0
.1
IV I II III IV I II III IV I II III IV I II III IV I
2009 2010 2011 2012 2013
D Residuals
-.10
-.05
.00
.05
.10
IV I II III IV I II III IV I II III IV I II III IV I
2009 2010 2011 2012 2013
E Residuals
-.15
-.10
-.05
.00
.05
.10
.15
IV I II III IV I II III IV I II III IV I II III IV I
2009 2010 2011 2012 2013
G Residuals
-.06
-.04
-.02
.00
.02
.04
.06
IV I II III IV I II III IV I II III IV I II III IV I
2009 2010 2011 2012 2013
H Residuals
-.08
-.04
.00
.04
.08
IV I II III IV I II III IV I II III IV I II III IV I
2009 2010 2011 2012 2013
I Residuals
-.20
-.15
-.10
-.05
.00
.05
.10
IV I II III IV I II III IV I II III IV I II III IV I
2009 2010 2011 2012 2013
J Residuals
-.10
-.05
.00
.05
.10
.15
IV I II III IV I II III IV I II III IV I II III IV I
2009 2010 2011 2012 2013
K Residuals
-.12
-.08
-.04
.00
.04
.08
IV I II III IV I II III IV I II III IV I II III IV I
2009 2010 2011 2012 2013
M Residuals
-.08
-.04
.00
.04
.08
.12
IV I II III IV I II III IV I II III IV I II III IV I
2009 2010 2011 2012 2013
O Residuals
-.08
-.04
.00
.04
.08
.12
IV I II III IV I II III IV I II III IV I II III IV I
2009 2010 2011 2012 2013
P Residuals
-.2
-.1
.0
.1
.2
.3
IV I II III IV I II III IV I II III IV I II III IV I
2009 2010 2011 2012 2013
Q Residuals
-.08
-.04
.00
.04
.08
.12
IV I II III IV I II III IV I II III IV I II III IV I
2009 2010 2011 2012 2013
R Residuals
Patikrinti modelio specifikaciją Omitted/Rdundnt variables test
Nereikalingo veiksnio tikrinimas
Patikrinti modelio specifikaciją Omitted/Rdundnt variables test
Ar dlogIMMA(-2)) perteklinis veiksnys
Redundant variables: DLOG(MMA(-2))
F-statistic 2.400141 Prob. F(1,251) 0.1226Log likelihood ratio 2.569559 Prob. Chi-Square(1) 0.1089
Test Equation:Dependent Variable: DLOG(UZ_?)Method: Panel Least SquaresDate: 12/10/13 Time: 23:48Sample: 2008Q4 2013Q1Included observations: 18Cross-sections included: 15Total pool (balanced) observations: 270
Variable Coefficient Std. Error t-Statistic Prob.
C -0.156110 0.018266 -8.546533 0.0000AUGIMAS_? 0.000814 0.000122 6.663143 0.0000@TREND 0.011276 0.002964 3.804194 0.0002
@TREND^2 -0.000368 0.000126 -2.915739 0.0039
Effects Specification
Cross-section fixed (dummy variables)
R-squared 0.295594 Mean dependent var -0.001930Adjusted R-squared 0.248074 S.D. dependent var 0.056786S.E. of regression 0.049241 Akaike info criterion -3.119848Sum squared resid 0.611013 Schwarz criterion -2.879953Log likelihood 439.1795 Hannan-Quinn criter. -3.023517F-statistic 6.220481 Durbin-Watson stat 2.104804Prob(F-statistic) 0.000000
Tikrinama ar trendas yra būtinas modeliui Redundant variables: @TREND @TREND^2
F-statistic 11.11106 Prob. F(2,282) 0.0000Log likelihood ratio 22.75531 Prob. Chi-Square(2) 0.0000
Test Equation:Dependent Variable: DLOG(UZ_?)Method: Panel Least SquaresDate: 12/10/13 Time: 23:53Sample: 2008Q2 2013Q1Included observations: 20Cross-sections included: 15Total pool (balanced) observations: 300
Variable Coefficient Std. Error t-Statistic Prob.
C -0.090036 0.012637 -7.125076 0.0000AUGIMAS_? 0.000856 0.000120 7.158023 0.0000
Effects Specification
Cross-section fixed (dummy variables)
R-squared 0.197128 Mean dependent var -0.001998Adjusted R-squared 0.154723 S.D. dependent var 0.054641S.E. of regression 0.050236 Akaike info criterion -3.092310Sum squared resid 0.716719 Schwarz criterion -2.894775Log likelihood 479.8465 Hannan-Quinn criter. -3.013256F-statistic 4.648681 Durbin-Watson stat 1.938516Prob(F-statistic) 0.000000
Atmetame hipotezę
Koeficientų interpretacija• LOG(UZ_A) = -0.0275779318163 - 0.121241681469 + 0.000858855763486*AUGIMAS_A +
0.00424616977558*@TREND - 9.56430278843e-05*@TREND^2
• DLOG(UZ_B) = -0.006235340038 - 0.121241681469 + 0.000858855763486*AUGIMAS_B + 0.00424616977558*@TREND - 9.56430278843e-05*@TREND^2
• DLOG(UZ_C) = -0.00650443893061 - 0.121241681469 + 0.000858855763486*AUGIMAS_C + 0.00424616977558*@TREND - 9.56430278843e-05*@TREND^2
• DLOG(UZ_D) = -0.0128625265146 - 0.121241681469 + 0.000858855763486*AUGIMAS_D + 0.00424616977558*@TREND - 9.56430278843e-05*@TREND^2
• DLOG(UZ_E) = 0.00961520809271 - 0.121241681469 + 0.000858855763486*AUGIMAS_E + 0.00424616977558*@TREND - 9.56430278843e-05*@TREND^2
• DLOG(UZ_G) = -0.0274648068859 - 0.121241681469 + 0.000858855763486*AUGIMAS_G + 0.00424616977558*@TREND - 9.56430278843e-05*@TREND^2
• DLOG(UZ_H) = -0.00159539355618 - 0.121241681469 + 0.000858855763486*AUGIMAS_H + 0.00424616977558*@TREND - 9.56430278843e-05*@TREND^2
• DLOG(UZ_I) = 0.0111884954027 - 0.121241681469 + 0.000858855763486*AUGIMAS_I + 0.00424616977558*@TREND - 9.56430278843e-05*@TREND^2
• DLOG(UZ_J) = -0.00213506936617 - 0.121241681469 + 0.000858855763486*AUGIMAS_J + 0.00424616977558*@TREND - 9.56430278843e-05*@TREND^2
• DLOG(UZ_K) = 0.00918540653087 - 0.121241681469 + 0.000858855763486*AUGIMAS_K + 0.00424616977558*@TREND - 9.56430278843e-05*@TREND^2
• DLOG(UZ_M) = -0.00736209698098 - 0.121241681469 + 0.000858855763486*AUGIMAS_M + 0.00424616977558*@TREND - 9.56430278843e-05*@TREND^2
• DLOG(UZ_O) = 0.00444257138038 - 0.121241681469 + 0.000858855763486*AUGIMAS_O + 0.00424616977558*@TREND - 9.56430278843e-05*@TREND^2
• DLOG(UZ_P) = 0.0114592116877 - 0.121241681469 + 0.000858855763486*AUGIMAS_P + 0.00424616977558*@TREND - 9.56430278843e-05*@TREND^2
• DLOG(UZ_Q) = 0.0196708838217 - 0.121241681469 + 0.000858855763486*AUGIMAS_Q + 0.00424616977558*@TREND - 9.56430278843e-05*@TREND^2
• DLOG(UZ_R) = 0.0261758271727 - 0.121241681469 + 0.000858855763486*AUGIMAS_R + 0.00424616977558*@TREND - 9.56430278843e-05*@TREND^2