EViews.com

Hi All

Can anyone help in three-stage least square specification problems of Panel data, Eviews 8
I used the following original model of Panel data of listed firm between 2009 and 2013, total observation is 540 no of firm is 108 :
Tobin's q = a + β1Manager Ownership("MO") + β2MO2 + β3MO3 + control variables (TOTAL ASSETS , LEV, TURNOVER) + ε {1}
Capital Structure("CS") = a + β1MO + β2MO2 + β3MO3 + control variables (CR5, ROA, TURNOVER) + ε {2)

I used the following 3SLS system:
MOt=c(1)+c(2)*(Tobin's qt) +c(3)*(ROAt)+c(4)*CSt+c(5)*(MO)t-1
inst (MO t-1) {3}
Tobin'sqt=c(6)+c(7)*(MOt)+c(8)*(MOt)^2+c(9)*(MOt)^3+c(10)*(CSt)+c(11)*log(TOTALASSETSt)
+c(12)*log(TURNOVERt)+c(13)*(Tobin's q t-1)
inst (Tobin's q t-1) {4}
CSt=c(14)+c(15)*(MOt)+c(16)*(MOt)^2+c(17)*(MOt)^3+c(18)*CR5t+c(19)*(ROAt)+c(20)*log(TURNOVEt)+c(21)*(Tobin's qt)+c(22)*(CS t-1)
inst (CS t-1) {5}

When I run Objects|New Objects|System|Estimate, it can not run and the eView show a message : "near singular matrix"

Any on can help the following :
1) is it due to my wrong setting?
2) how to solve "near singular matrix"?
3) any wrong in variable? Do I need to delete variable to fix the problem? If yes, how to choose the variable need to deleted?

Attached descriptive analysis for information.