Statistic tests return "NA" in a FIML model.

[lcfreitas]

Hello,
I have been working in a system of equations using a FIML model to solve it in eviews. My sample consists of 23 years observations and my aim is to check the elasticity of substitution between fuels in a given economy. Output of the model using Marquardt optimization is as follows.

My doubt is: Why the output of the estimations does not show values for the statistics tests (Std. Error, S-Stat., Prob)?

Thank you in advance


System: SYS_1
Estimation Method: Full Information Maximum Likelihood (Marquardt)
Date: 02/26/13 Time: 20:10
Sample: 1989 2011
Included observations: 23
Total system (balanced) observations 92
Estimation settings: tol=0.00010, derivs=accurate numeric
Initial Values: C(11)=-2.92985, C(15)=-4.90309, C(1)=-1.62635, C(2)=
-2.58175, C(3)=-0.96767, C(4)=-0.61110, C(5)=-0.47247, C(6)=
-0.92587, C(16)=-0.25547, C(20)=0.84160, C(12)=-3.89147, C(7)=
-4.06322, C(8)=-0.88064, C(10)=-0.82580, C(17)=-0.13385, C(13)=
-7.20022, C(9)=-1.30223, C(18)=0.23618, C(14)=-4.50539, C(19)=
-0.04686
Convergence achieved after 1 iteration
WARNING: Singular covariance - coefficients are not unique

Coefficient Std. Error z-Statistic Prob.

C(11) -2.929853 NA NA NA
C(15) -4.903091 NA NA NA
C(1) -1.626346 NA NA NA
C(2) -2.581752 NA NA NA
C(3) -0.967672 NA NA NA
C(4) -0.611098 NA NA NA
C(5) -0.472469 NA NA NA
C(6) -0.925875 NA NA NA
C(16) -0.255470 NA NA NA
C(20) 0.841598 NA NA NA
C(12) -3.891465 NA NA NA
C(7) -4.063226 NA NA NA
C(8) -0.880638 NA NA NA
C(10) -0.825795 NA NA NA
C(17) -0.133848 NA NA NA
C(13) -7.200220 NA NA NA
C(9) -1.302233 NA NA NA
C(18) 0.236178 NA NA NA
C(14) -4.505393 NA NA NA
C(19) -0.046860 NA NA NA

Log likelihood 373.7117 Schwarz criterion -29.77016
Avg. log likelihood 4.062084 Hannan-Quinn criter. -30.50922
Akaike info criterion -30.75754
Determinant residual covariance 5.62E-11


Equation: LOG(WET / WOI) = (C(11) - C(15)) - (((C(1) * WBI) + (C(2) * WNG)
+ (C(3) * WEL) + (C(4) * (WET + WOI))) * LOG(PET / POI)) + (C(1) - C(5))
* WBI * LOG(PBI / POI) + (C(2) - C(5)) * WNG * LOG(PNG / POI) + (C(3) -
C(6)) * WEL * LOG(PEL / POI) + C(16) * LOG(GDPPC(-1)) + C(20) *
LOG(CET(-1) / COI(-1))
Observations: 23
R-squared 0.884195 Mean dependent var -1.523633
Adjusted R-squared 0.804023 S.D. dependent var 0.290054
S.E. of regression 0.128405 Sum squared resid 0.214341
Durbin-Watson stat 1.457975

Equation: LOG(WBI / WOI) = (C(12) - C(15)) - (((C(1) * WET) + (C(7) * WNG)
+ (C(8) * WEL) + (C(10) * (WBI + WOI))) * LOG(PBI / POI)) + (C(1) -
C(10)) * WET * LOG(PET / POI) + (C(7) - C(5)) * WNG * LOG(PNG / POI)
+ (C(8) - C(6)) * WEL * LOG(PEL / POI) + C(17) * LOG(GDPPC(-1)) +
C(20) * LOG(CBI(-1) / COI(-1))
Observations: 23
R-squared 0.935352 Mean dependent var -1.854431
Adjusted R-squared 0.890595 S.D. dependent var 0.181722
S.E. of regression 0.060107 Sum squared resid 0.046967
Durbin-Watson stat 2.285238

Equation: LOG(WNG / WOI) = (C(13) - C(15)) - (((C(2) * WET) + (C(7) *
WNG) + (C(9) * WEL) + (C(5) * (WNG + WOI))) * LOG(PNG / POI)) +
(C(2) - C(4)) * WET * LOG(PET / POI) + (C(7) - C(10)) * WNG *
LOG(PNG / POI) + (C(9) - C(6)) * WEL * LOG(PEL / POI) + C(18) *
LOG(GDPPC(-1)) + C(20) * LOG(CNG(-1) / COI(-1))
Observations: 23
R-squared 0.988806 Mean dependent var -3.267752
Adjusted R-squared 0.979477 S.D. dependent var 0.621961
S.E. of regression 0.089100 Sum squared resid 0.095266
Durbin-Watson stat 2.015790

Equation: LOG(WEL / WOI) = (C(14) - C(15)) - (((C(3) * WET) + (C(8) * WBI)
+ (C(9) * WNG) + (C(6) * (WEL + WOI))) * LOG(PEL / POI)) + (C(3) -
C(4)) * WET * LOG(PET / POI) + (C(8) - C(10)) * WBI * LOG(PBI / POI) +
(C(9) - C(5)) * WNG * LOG(PNG / POI) + C(19) * LOG(GDPPC(-1)) +
C(20) * LOG(CEL(-1) / COI(-1))
Observations: 23
R-squared 0.930435 Mean dependent var 0.086509
Adjusted R-squared 0.872465 S.D. dependent var 0.133160
S.E. of regression 0.047554 Sum squared resid 0.027137
Durbin-Watson stat 1.599228

[EViews Gareth]

WARNING: Singular covariance - coefficients are not unique