How to report intercept iterm in FixedEffectModels?

General Usage
1

Dear all,

In FixedEffectModels with fe, the result doesn’t report intercept term. Anyone knows how to report constant term?

julia> df = dataset("plm", "GrunFeld")
julia> reg(df, @formula(Inv ~ Value + Capital + Year + fe(Firm)))
                        Fixed Effect Model
==================================================================
Number of obs:               200  Degrees of freedom:           13
R2:                        0.947  R2 Adjusted:               0.943
F-Stat:                  219.159  p-value:                   0.000
R2 within:                 0.779  Iterations:                    1
==================================================================
Inv     | Estimate Std.Error  t value Pr(>|t|) Lower 95% Upper 95%
------------------------------------------------------------------
Value   | 0.110721 0.0115852   9.5571    0.000 0.0878662  0.133575
Capital | 0.353577 0.0218494  16.1824    0.000  0.310474   0.39668
Year    | -2.66422  0.843852 -3.15721    0.002  -4.32891 -0.999525
==================================================================

In Stata, it always report intercept term.

. xtreg inv mv ks c.year, fe

Fixed-effects (within) regression               Number of obs     =        200
Group variable: company                         Number of groups  =         10

R-squared:                                      Obs per group:
     Within  = 0.7786                                         min =         20
     Between = 0.8108                                         avg =       20.0
     Overall = 0.8010                                         max =         20

                                                F(3,187)          =     219.16
corr(u_i, Xb) = -0.2367                         Prob > F          =     0.0000

------------------------------------------------------------------------------
      invest | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
      mvalue |     0.1107     0.0116     9.56   0.000       0.0879      0.1336
      kstock |     0.3536     0.0218    16.18   0.000       0.3105      0.3967
        year |    -2.6642     0.8439    -3.16   0.002      -4.3289     -0.9995
       _cons |  5109.1716  1636.9071     3.12   0.002    1879.9942   8338.3489
-------------+----------------------------------------------------------------
     sigma_u |   89.64192
     sigma_e |  51.552705
         rho |  .75146422   (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(9, 187) = 52.51                     Prob > F = 0.0000

In econometrics, we often need to report intercept term.

2

I’d say Stata is the odd one out here, it’s common not to report an intercept in fixed effect models (as it would be collinear with the fixed effects).

See here for what Stata does, you should be able to replicate this with FixedEffectModels:

3

I try to use Econometrics, Then I can get the constant term. But it can not be reported by RegressionTables.

julia> using Econometrics
[ Info: Precompiling Econometrics [4d6a76a9-bfbc-5492-8924-cf6ed7875f06]

help?> Econometrics
search: Econometrics EconometricsMLJModel EconometricModel

  Econometrics

  Econometrics in Julia.


julia> model = fit(EconometricModel, @formula(Inv ~ Value + Capital + Year + absorb(Firm)), df)
Continuous Response Model
Number of observations: 200
Null Loglikelihood: -1359.15
Loglikelihood: -1065.81
R-squared: 0.9469
Wald: 219.16 ∼ F(3, 187) ⟹ Pr > F = 0.0000
Formula: Inv ~ 1 + Value + Capital + Year + absorb(Firm)
Variance Covariance Estimator: OIM
─────────────────────────────────────────────────────────────────────────────────────
                   PE            SE      t-value  Pr > |t|         2.50%       97.50%
─────────────────────────────────────────────────────────────────────────────────────
(Intercept)  5109.17      1636.91        3.12124    0.0021  1879.99       8338.35
Value           0.110721     0.0115852   9.5571     <1e-17     0.0878662     0.133575
Capital         0.353577     0.0218494  16.1824     <1e-36     0.310474      0.39668
Year           -2.66422      0.843852   -3.15721    0.0019    -4.32891      -0.999525
─────────────────────────────────────────────────────────────────────────────────────

julia> using RegressionTables

julia> regtable(model)
ERROR: MethodError: Cannot `convert` an object of type Vector{String} to an object of type String
Closest candidates are:
  convert(::Type{S}, ::CategoricalValue) where S<:Union{AbstractChar, AbstractString, Number} at ~/.julia/packages/CategoricalArrays/Fr04b/src/value.jl:79
  convert(::Type{String}, ::FilePathsBase.AbstractPath) at ~/.julia/packages/FilePathsBase/dVWLK/src/path.jl:117
  convert(::Type{String}, ::WeakRefStrings.WeakRefString) at ~/.julia/packages/WeakRefStrings/31nkb/src/WeakRefStrings.jl:81
4

But Econometrics.jl also report intercept term.