LSDV3 model: dummy variable omitted because of collinearity

[Giuseppe Polito]

I am studying the notes titled "Practical Guides To Panel Data Modeling: A Step by Step Analysis Using Stata*" by Hun Myoung Park (http://www.iuj.ac.jp/faculty/kucc625.../panel_iuj.pdf)

In particular I am trying to fit a LSDV3 model. This model includes the intercept and all dummies, and then impose a restriction that the sum of parameters of all dummies is zero. As explained in the notes, the command that I use is

constraint define 1 g1+g2+g3+g4+g5+g6=0
cnsreg cost g1-g6 output fuel load, constraint(1)

but STATA nevertheless drops 1 dummy variable (it should not be supposed to do that, since the model is identified through the constraint). In particular cnsreg gives me

note: g6 omitted because of collinearity

How I can replicate the LSDV3 model explained in the paper?

I have STATA version 14.2

[Phil Bromiley]

Welcome to Stata list. You will increase your chances of useful answer by following the FAQ on asking questions – provide Stata code in code delimiters which you do, readable Stata output, and sample data in dataex. Why the code appeared to work in his implementation and not yours, I have no idea. From a practical standpoint, there are several other ways to estimate precisely this model and they are explained in the paper you cite.