Page 1 of 1

Length of data for Bayesian estimation

PostPosted: Sat Jul 06, 2013 1:42 pm
by richardgu26
Dear all,

I have a quick question: in Bayesian estimation, mapping data into model, we are supposed to have a series of data. However, if the length of certain series is less than others, will estimation be affected? The reason I am asking is that when I am following up Chen et al (2012) "Macro effect of LSAP Program", I have to find short-term bills to long-term bonds ratio, e.g. outstanding T-bills versus T-bonds. In appendix, it says all data can be extracted from FRED St. Louis. And what I can find is long-term and short-term treasury securities from 2002 to present.

I don't think that works in Bayesian estimation, right?

Thanks!

Re: Length of data for Bayesian estimation

PostPosted: Sat Jul 06, 2013 3:59 pm
by jpfeifer
You can probably treat those as missing values in the Kalman filter.

Re: Length of data for Bayesian estimation

PostPosted: Sun Jul 07, 2013 12:41 am
by richardgu26
jpfeifer wrote:You can probably treat those as missing values in the Kalman filter.



@jpfeifer

Thanks for your quick reply. I will try your suggestion see if it worked.