Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

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

Re: st: out-of-sample prediction does not work after -regress- with lags

From   maarten buis <>
Subject   Re: st: out-of-sample prediction does not work after -regress- with lags
Date   Wed, 4 May 2011 17:26:21 +0200

On Wed, May 4, 2011 at 4:50 PM, Mirko <> wrote:
> I am not interested in the missing values at the beginng of the
> sample, clearly they are missing because there is not enough
> information. Instead, I would like to undestand why -predict- perfoms
> the out-of-sample forecast for one data point at the end of the series
> (the observations created with -tsappend-) only and not for the whole
> sample when lags are presents.

For your out of sample forcast you added five observations at the end.
By adding 5 extra years you did not also include the unemployment
rate for those years. After all, how could Stata know those? So the
unemployment rate for these years is missing, and Stata does not
know what to predict for these years.

Out of sample prediction is not black/white magic, it is just the filling
in of a formula. Your data must have information on all explanatory
variables to do an out of sample prediction, so just typing -tsappend-
will not do that (unless time is your only explanatory variable).

Hope this helps,

Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index