problem with forecast AR(1) ar(6)

[avera]

I am trying to estimate an equation for sample from 2005-jan to 2013-nov in a single equation ar(1) ar(6) ma(1) however the forcast for the variable is the same from 2009-jan to 2014. I tried with holdrick prescot as well and I get the same problem after some time 2006-04 there is not change in the forecast, what can be happened same thing is happennig with all variables. thank you

[EViews Gareth]

Without knowing exactly what you're doing, it is hard to say what you're doing wrong.

[avera]

Hi thank you for your answer,

I have the following data:

date v
01/01/2005 0,82%
01/02/2005 1,02%
01/03/2005 0,77%
01/04/2005 0,44%
01/05/2005 0,41%
01/06/2005 0,40%
01/07/2005 0,05%
01/08/2005 0,00%
01/09/2005 0,43%
01/10/2005 0,23%
01/11/2005 0,11%
01/12/2005 0,07%
01/01/2006 0,54%
01/02/2006 0,66%
01/03/2006 0,70%
01/04/2006 0,45%
01/05/2006 0,33%
01/06/2006 0,30%
01/07/2006 0,41%
01/08/2006 0,39%
01/09/2006 0,29%
01/10/2006 -0,14%
01/11/2006 0,24%
01/12/2006 0,23%
01/01/2007 0,77%
01/02/2007 1,17%
01/03/2007 1,21%
01/04/2007 0,90%
01/05/2007 0,30%
01/06/2007 0,12%
01/07/2007 0,17%
01/08/2007 -0,13%
01/09/2007 0,08%
01/10/2007 0,01%
01/11/2007 0,47%
01/12/2007 0,49%
01/01/2008 1,06%
01/02/2008 1,51%
01/03/2008 0,81%
01/04/2008 0,71%
01/05/2008 0,93%
01/06/2008 0,86%
01/07/2008 0,48%
01/08/2008 0,19%
01/09/2008 -0,19%
01/10/2008 0,35%
01/11/2008 0,28%
01/12/2008 0,44%
01/01/2009 0,59%
01/02/2009 0,84%
01/03/2009 0,50%
01/04/2009 0,32%
01/05/2009 0,01%
01/06/2009 -0,06%
01/07/2009 -0,04%
01/08/2009 0,04%
01/09/2009 -0,11%
01/10/2009 -0,13%
01/11/2009 -0,07%
01/12/2009 0,08%
01/01/2010 0,69%
01/02/2010 0,83%
01/03/2010 0,25%
01/04/2010 0,46%
01/05/2010 0,10%
01/06/2010 0,11%
01/07/2010 -0,04%
01/08/2010 0,11%
01/09/2010 -0,14%
01/10/2010 -0,09%
01/11/2010 0,19%
01/12/2010 0,65%
01/01/2011 0,91%
01/02/2011 0,60%
01/03/2011 0,27%
01/04/2011 0,12%
01/05/2011 0,28%
01/06/2011 0,32%
01/07/2011 0,14%
01/08/2011 -0,03%
01/09/2011 0,31%
01/10/2011 0,19%
01/11/2011 0,14%
01/12/2011 0,42%
01/01/2012 0,73%
01/02/2012 0,61%
01/03/2012 0,12%
01/04/2012 0,14%
01/05/2012 0,30%
01/06/2012 0,08%
01/07/2012 -0,02%
01/08/2012 0,04%
01/09/2012 0,29%
01/10/2012 0,16%
01/11/2012 -0,14%
01/12/2012 0,09%
01/01/2013 0,30%
01/02/2013 0,44%
01/03/2013 0,21%
01/04/2013 0,25%
01/05/2013 0,28%
01/06/2013 0,23%
01/07/2013 0,04%
01/08/2013 0,08%
01/09/2013 0,29%
01/10/2013 -0,26%
01/11/2013 -0,22%

however when I estimate the equation ar(1) ar (6) the results I get in the forecast are those:
DATE VFORECAST
01/01/2005 NA
01/02/2005 NA
01/03/2005 NA
01/04/2005 NA
01/05/2005 NA
01/06/2005 NA
01/07/2005 0,253%
01/08/2005 0,156%
01/09/2005 0,152%
01/10/2005 0,209%
01/11/2005 0,243%
01/12/2005 0,261%
01/01/2006 0,297%
01/02/2006 0,333%
01/03/2006 0,351%
01/04/2006 0,350%
01/05/2006 0,344%
01/06/2006 0,337%
01/07/2006 0,327%
01/08/2006 0,316%
01/09/2006 0,307%
01/10/2006 0,303%
01/11/2006 0,302%
01/12/2006 0,302%
01/01/2007 0,304%
01/02/2007 0,308%
01/03/2007 0,311%
01/04/2007 0,313%
01/05/2007 0,314%
01/06/2007 0,315%
01/07/2007 0,315%
01/08/2007 0,314%
01/09/2007 0,313%
01/10/2007 0,313%
01/11/2007 0,312%
01/12/2007 0,311%
01/01/2008 0,311%
01/02/2008 0,311%
01/03/2008 0,311%
01/04/2008 0,312%
01/05/2008 0,312%
01/06/2008 0,312%
01/07/2008 0,312%
01/08/2008 0,312%
01/09/2008 0,312%
01/10/2008 0,312%
01/11/2008 0,312%
01/12/2008 0,312%
01/01/2009 0,312%
01/02/2009 0,312%
01/03/2009 0,312%
01/04/2009 0,312%
01/05/2009 0,312%
01/06/2009 0,312%
01/07/2009 0,312%
01/08/2009 0,312%
01/09/2009 0,312%
01/10/2009 0,312%
01/11/2009 0,312%
01/12/2009 0,312%
01/01/2010 0,312%
01/02/2010 0,312%
01/03/2010 0,312%
01/04/2010 0,312%
01/05/2010 0,312%
01/06/2010 0,312%
01/07/2010 0,312%
01/08/2010 0,312%
01/09/2010 0,312%
01/10/2010 0,312%
01/11/2010 0,312%
01/12/2010 0,312%
01/01/2011 0,312%
01/02/2011 0,312%
01/03/2011 0,312%
01/04/2011 0,312%
01/05/2011 0,312%
01/06/2011 0,312%
01/07/2011 0,312%
01/08/2011 0,312%
01/09/2011 0,312%
01/10/2011 0,312%
01/11/2011 0,312%
01/12/2011 0,312%
01/01/2012 0,312%
01/02/2012 0,312%
01/03/2012 0,312%
01/04/2012 0,312%
01/05/2012 0,312%
01/06/2012 0,312%
01/07/2012 0,312%
01/08/2012 0,312%
01/09/2012 0,312%
01/10/2012 0,312%
01/11/2012 0,312%
01/12/2012 0,312%
01/01/2013 0,312%
01/02/2013 0,312%
01/03/2013 0,312%
01/04/2013 0,312%
01/05/2013 0,312%
01/06/2013 0,312%
01/07/2013 0,312%
01/08/2013 0,312%
01/09/2013 0,312%
01/10/2013 0,312%
01/11/2013 0,312%
01/12/2013 0,312%
01/01/2014 0,312%
01/02/2014 0,312%
01/03/2014 0,312%
01/04/2014 0,312%
01/05/2014 0,312%
01/06/2014 0,312%
01/07/2014 0,312%
01/08/2014 0,312%
01/09/2014 0,312%
01/10/2014 0,312%
01/11/2014 0,312%
01/12/2014 0,312%

I dont know why it is not changing the values after some point, also I make a holdrick presccot filter to the original data, and I tried to forecast with the adjusted data but I got same problem after some data points.

thank you for your help

[startz]

What is the version and build date of your copy of EViews? (Help/About EViews)
What is the smpl set at? (It would be better if you posted the workfile including the equation you used for the forecast.)

[trubador]

I dont know why it is not changing the values after some point, also I make a holdrick presccot filter to the original data, and I tried to forecast with the adjusted data but I got same problem after some data points.

I think there is a confusion between the one-step ahead and multi-step ahead forecasting exercise here. If former is the case, you can choose "Static Forecast" to see how your model fits your data in-sample (e.g. Jan 2005-Nov 2013). In the latter case, "dynamic forecast" would be the appropriate choice (e.g. Dec 2013 - Dec 2014). I can tell from the results that you have carried out an in-sample dynamic forecasting. And therefore, forecasts have converged to their conditional mean at some point, as expected...

[avera]

I dont know why it is not changing the values after some point, also I make a holdrick presccot filter to the original data, and I tried to forecast with the adjusted data but I got same problem after some data points.

I think there is a confusion between the one-step ahead and multi-step ahead forecasting exercise here. If former is the case, you can choose "Static Forecast" to see how your model fits your data in-sample (e.g. Jan 2005-Nov 2013). In the latter case, "dynamic forecast" would be the appropriate choice (e.g. Dec 2013 - Dec 2014). I can tell from the results that you have carried out an in-sample dynamic forecasting. And therefore, forecasts have converged to their conditional mean at some point, as expected...

Thank you, your are right initially I used the dynamic forecast, but then when I change it to static I can only get forecast for 1 month dec 2013, Do you any suggestions in how can I get the results the other months 2014-jan to 2014-dec?

[trubador]

Select "Dynamic forecast" and set the sample to "2013M12 2014M12". If you are not familiar with ARMA forecasting, please refer to a textbook or visit the following discussion in the forum: ARIMA Estimation and Forecasting As for the EViews implementation of forecasting in general, I also recommend the Chapter 8 of Startz' "Eviews Illustrated", which also has a very nice and brief introduction of the theory.

[avera]

thank you it solves the problem! :D