Bayesian VARs in combination with the kalman filter
Posted: Sun Aug 31, 2014 1:13 pm
Dear Sir or Madam,
i have two questions concerning the above mentioned topic:
1. What is the easiest way to estimate a random coefficient bvar in the kalman filter setting? Perhaps the same question in other words: Can i use the bvar estimation framework in eviews within the kalman filter setup?
2. In my finalized master thesis, i had programmed a recursive forecasting scheme for a ordinary var model (see topic from Aug 27, 2012). Therefore, my target is to get a time series of - for example - one-step-ahead-forecasts (the corresponding forecast origin will be shifted equally one-step-ahead, as a new data point becomes available). But neither the dynamic nor the smoothed forecast function in eviews seems to do the same as my programmed recursive forecasting scheme. How can i use the kalman filter for constructing a recursive forecasting scheme?
I'm delighted to see your answers. Thank you!
Goodnight!
Yohan.
i have two questions concerning the above mentioned topic:
1. What is the easiest way to estimate a random coefficient bvar in the kalman filter setting? Perhaps the same question in other words: Can i use the bvar estimation framework in eviews within the kalman filter setup?
2. In my finalized master thesis, i had programmed a recursive forecasting scheme for a ordinary var model (see topic from Aug 27, 2012). Therefore, my target is to get a time series of - for example - one-step-ahead-forecasts (the corresponding forecast origin will be shifted equally one-step-ahead, as a new data point becomes available). But neither the dynamic nor the smoothed forecast function in eviews seems to do the same as my programmed recursive forecasting scheme. How can i use the kalman filter for constructing a recursive forecasting scheme?
I'm delighted to see your answers. Thank you!
Goodnight!
Yohan.