[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

From |
"Sebastian Kruk" <residuo.solow@gmail.com> |

To |
statalist@hsphsun2.harvard.edu |

Subject |
st: complex VAR for preselection of instruments |

Date |
Thu, 18 Dec 2008 20:20:20 -0200 |

I would like to do a preselection of possible instruments within a VAR. The two endogenous variables variables π(t+1) and s(t) are regressed on all potential instruments. This specification can be formalized as: π(t+1)=v1+A11*π(t-1)+A12*π(t-2)+...+A1j*π(t-j)+B11*x(t-1)+B12*x(t-2)+...+B1j*x(t-j) s(t)=v2+A21*s(t-1)+A22*s(t-2)+...+A2j*π(t-j)+B21*x(t-1)+B22*x(t-2)+...+B2j*x(t-j) with v1,v2 a a deterministic term, and x(t−j) the vector of all other predetermined variables with lag j. The maximal lag length is L1 = 10 and L2 = 5. Can I apply a model reduction procedure that works through a sequential elimination of regressors in order to obtain a model that lead to the smallest value of the particular information criterion as Lutkepohl, H. (2005). New introduction to multiple time series analysis, Springer? Thanks in advance, Sebastian. * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/

- Prev by Date:
**RE: st: Why no predicted probs but exponentiated linear prediction for -adjust- after -xtgee- or -xtmelogit-?** - Next by Date:
**Re: st: Constant terms in AR1 error regressions** - Previous by thread:
**st: Dropped variabled in Simultaneuos Equations** - Next by thread:
**st: display width in tabulate** - Index(es):

© Copyright 1996–2015 StataCorp LP | Terms of use | Privacy | Contact us | What's new | Site index |