## Panel PMG/ARDL

For technical questions regarding estimation of single equations, systems, VARs, Factor analysis and State Space Models in EViews. General econometric questions and advice should go in the Econometric Discussions forum.

Moderators: EViews Gareth, EViews Moderator

mchan
Posts: 4
Joined: Tue Sep 19, 2017 3:04 am

### Panel PMG/ARDL

Hi all.

I am using Eviews 9 to estimate a PMG/ARDL model using unbalanced panel data, and there are a couple of clarifications I need to make. My estimation output window has both 'Long Run Equation' and 'Short Run Equation'.

My inputed dependent is LGDP and a ARDL (2,1,1,1...) model was selected:

1. How do I interpret the coefficients of the Long Run Equation – can my dependent be interpreted as LGDP or D(LGDP)? i.e. LGDP or D(GDP)= c(1)*x1+ C(2)*x2+ C(3)*x3+… + C(n)*xn? My confusion stems from the dependent variable being stated to be D(LGDP) in my output window.

2. Is it possible to generate the estimated Y based on solely the LR cointegrating relationship? The proc > forecast function only gives estimates for the whole ARDL model on a dynamic/ static forecast basis.

3. How do we conduct the bounds testing to test panel PMG/ARDL? Under coefficient diagnostics, only the Wald test is available- can I use to test for joint significance for my cointegrating coefficients and compare my F-statistic with the F-tables on Pesaran et al. (2001)?

4. Is there a function to test for heteroskedasticity of residuals? Residual diagnostics only have a histogram- normality test option

Thank you!

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 11765
Joined: Tue Sep 16, 2008 5:38 pm

### Re: Panel PMG/ARDL

2. Not built in
3. Bounds test is not appropriate for panels.
4. No.

mchan
Posts: 4
Joined: Tue Sep 19, 2017 3:04 am

### Re: Panel PMG/ARDL

Thanks for the reply Gareth, really apprecaite it.

How about my first question, should my dependent be interpreted as LGDP or D(LGDP)?

Cheers

mchan
Posts: 4
Joined: Tue Sep 19, 2017 3:04 am

### Re: Panel PMG/ARDL

Does anyone knows if we can obtain an estimate for our dependent using the coefficients from only the Long-Run Equation coefficients? I tried doing this on eviews (i.e. LGDP_Forecast= c(1)*x1 + c(2)*x2 +.. c(n)*xn) but got a very inflated estimate (100 times my observed) and does not co-move at all.

Wondering if I intepreted the coefficients in my long-run equation wrongly.

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 11765
Joined: Tue Sep 16, 2008 5:38 pm

### Re: Panel PMG/ARDL

Without going into the econometrics, the equation is estimated in differences.