[Gretl-users] ARMAX question

[lisa paul]

Could someone explain to me the last bit of the ARMAX output?  I would think it 
means that there's a unit root (or explosive), but when I run an augmented 
dickey-fuller test, it says no unit root...thanks!


Model 1:
ARMAX, using observations 1957:06-1979:08 (T = 267)
Estimated
using Kalman filter (exact ML)
Dependent
variable: Ratio
Standard
errors based on Hessian
 
             coefficient   std. error      z       p-value 
  ---------------------------------------------------------
  const      0.650441      0.0246480     26.39    1.83e-153 ***
  phi_1      0.876993      0.0297168     29.51    2.04e-191 ***
  var23   0.00235627    0.000457168    5.154   2.55e-07  ***
 
Mean
dependent var   0.735553   S.D. dependent var   0.083849
Mean of
innovations -0.000334   S.D. of
innovations  0.035214
Log-likelihood       513.8762   Akaike criterion    -1019.752
Schwarz
criterion   -1005.403   Hannan-Quinn        -1013.989
 
               Real  Imaginary    Modulus  Frequency
  -----------------------------------------------------------
  AR Root  1   1.1403  0.0000     1.1403     0.0000
  -----------------------------------------------------------



      

Could someone explain to me the last bit of the ARMAX output?  I would think it means that there's a unit root (or explosive), but when I run an augmented dickey-fuller test, it says no unit root...thanks!

Model 1: ARMAX, using observations 1957:06-1979:08 (T = 267)

Estimated using Kalman filter (exact ML)

Dependent variable: Ratio

Standard errors based on Hessian

 

             coefficient   std. error      z       p-value

  ---------------------------------------------------------

  const      0.650441      0.0246480     26.39    1.83e-153 ***

  phi_1      0.876993      0.0297168     29.51    2.04e-191 ***

  var23    0.00235627    0.000457168    5.154   2.55e-07  ***

 

Mean dependent var   0.735553   S.D. dependent var   0.083849

Mean of innovations -0.000334   S.D. of innovations  0.035214

Log-likelihood       513.8762   Akaike criterion    -1019.752

Schwarz criterion   -1005.403   Hannan-Quinn        -1013.989

 

               Real  Imaginary    Modulus  Frequency

  -----------------------------------------------------------

  AR Root  1   1.1403  0.0000     1.1403     0.0000

  -----------------------------------------------------------