Misspecification test after Fixed effects panel regression

[Radhika Channanamchery]

Hi, when I run the below codes along with the FE panel regression model to see if there is any misspecification, In the results, I see residuals are omitted due to collinearity. Could someone help me figure out why or how I can fix this? Thank you in advance

xtreg shareveggca shareofscandst share_urbanpop shareofirr_final populationdensitypersqkm gddplog roadper100km numberofbanksper1000sqkm
> rainfallcm temp shareofmarginalandsmall fitted4 sq_fitted4,fe vce(cluster disid)
note: fitted4 omitted because of collinearity

Fixed-effects (within) regression Number of obs = 135
Group variable: disid Number of groups = 27

R-sq: within = 0.1655 Obs per group: min = 5
between = 0.0066 avg = 5.0
overall = 0.0016 max = 5

F(11,26) = 7.56
corr(u_i, Xb) = -0.7697 Prob > F = 0.0000

(Std. Err. adjusted for 27 clusters in disid)
------------------------------------------------------------------------------------------
| Robust
shareveggca | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------------------+----------------------------------------------------------------
shareofscandst | .1284497 .1391531 0.92 0.364 -.1575835 .4144829
share_urbanpop | -.1960408 .1066919 -1.84 0.078 -.415349 .0232675
shareofirr_final | -.0014623 .002667 -0.55 0.588 -.0069443 .0040197
populationdensitypersqkm | -.0017243 .0011705 -1.47 0.153 -.0041303 .0006817
gddplog | .7203003 .4773254 1.51 0.143 -.2608561 1.701457
roadper100km | -.0467038 .0432493 -1.08 0.290 -.135604 .0421965
numberofbanksper1000sqkm | -.0078059 .0142947 -0.55 0.590 -.0371891 .0215773
rainfallcm | .0031424 .0094316 0.33 0.742 -.0162446 .0225293
temp | .0937392 .2056783 0.46 0.652 -.3290385 .516517
shareofmarginalandsmall | .0564681 .0607721 0.93 0.361 -.0684508 .181387
fitted4 | 0 (omitted)
sq_fitted4 | .022508 .0511524 0.44 0.664 -.0826373 .1276534
_cons | -3.748309 7.069376 -0.53 0.600 -18.27962 10.783
-------------------------+----------------------------------------------------------------
sigma_u | 5.023149
sigma_e | 1.6528293
rho | .90230806 (fraction of variance due to u_i)
------------------------------------------------------------------------------------------

. test sq_fitted4

( 1) sq_fitted4 = 0

F( 1, 26) = 0.19
Prob > F = 0.6636

.

Radhika C

[Clyde Schechter]

One possibility comes quickly to mind: if fitted4 is a dichotomous variable and sq_fitted4 is its square, then they will necessarily be colinear with each other. But, frankly, without example data to work with, any suggestion, including this, is just speculation. If my guess here is wrong, do show example data (including all the variables in the regression command) when posting back. Be sure to use the -dataex- command to do that, so that responders can work with your data, not just look at it.

If you are running version 18, 17, 16 or a fully updated version 15.1 or 14.2, -dataex- is already part of your official Stata installation. If not, run -ssc install dataex- to get it. Either way, run -help dataex- to read the simple instructions for using it. -dataex- will save you time; it is easier and quicker than typing out tables. It includes complete information about aspects of the data that are often critical to answering your question but cannot be seen from tabular displays or screenshots. It also makes it possible for those who want to help you to create a faithful representation of your example to try out their code, which in turn makes it more likely that their answer will actually work in your data.

[George Ford]

You're running a Ramsey Reset test. The fitted values should all be polynomials, typically yfit^2 and yfit^3

[Jeff Wooldridge]

The fitted values are linear function of the explanatory variables, so there is always perfect collinearity. As George says, you should include the square and cube. And this is purely a functional form test.

You don’t show how you obtained the fitted values, but they should not depend on the estimated fixed effects.

[Radhika Channanamchery]

You're running a Ramsey Reset test. The fitted values should all be polynomials, typically yfit^2 and yfit^3

Thank you for your kind reply


Radhika

[Radhika Channanamchery]

The fitted values are linear function of the explanatory variables, so there is always perfect collinearity. As George says, you should include the square and cube. And this is purely a functional form test.

You don’t show how you obtained the fitted values, but they should not depend on the estimated fixed effects.

Thank you for your kind reply Professor Wooldridge

Radhika

[Radhika Channanamchery]

One possibility comes quickly to mind: if fitted4 is a dichotomous variable and sq_fitted4 is its square, then they will necessarily be colinear with each other. But, frankly, without example data to work with, any suggestion, including this, is just speculation. If my guess here is wrong, do show example data (including all the variables in the regression command) when posting back. Be sure to use the -dataex- command to do that, so that responders can work with your data, not just look at it.

If you are running version 18, 17, 16 or a fully updated version 15.1 or 14.2, -dataex- is already part of your official Stata installation. If not, run -ssc install dataex- to get it. Either way, run -help dataex- to read the simple instructions for using it. -dataex- will save you time; it is easier and quicker than typing out tables. It includes complete information about aspects of the data that are often critical to answering your question but cannot be seen from tabular displays or screenshots. It also makes it possible for those who want to help you to create a faithful representation of your example to try out their code, which in turn makes it more likely that their answer will actually work in your data.

Thanks for your reply. I will use dataex in future posts, and I appreciate your suggestion.