[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: re: temporal disaggregation

From   Kit Baum <[email protected]>
To   [email protected]
Subject   st: re: temporal disaggregation
Date   Mon, 3 Sep 2007 21:40:01 -0400

Michael wrote

> I have a yearly gross product data and I would like to transform to
> montly data.
> Is there a program to do it?

You may be looking for something like the Chow-Lin procedure.
(Google it, and/or "disaggregation".) It uses higher-frequency
series that are conceptually correlated with your lower-frequency
series (e.g. GDP) to make imputations about the hypothetical higher-
frequency values of that series. I know a procedure to do so has
been implemented in RATS; I have seen papers reference Gauss
programs that implement it; to the best of my knowledge it has not
been programmed in Stata. There is a literature on this topic that
you should explore, and some authors have suggested alternatives.

The World Bank's preferred method is the Denton procedure; findit denton within Stata. That only alters the frequency by one, though-- from annual to quarterly, or from quarterly to monthly. I agree with Michael that going from annual to monthly is likely to be befraught with problems. I would recommend using an available monthly series as a proxy, ignoring the GDP figures. A popular choice is an industrial production index.

Kit Baum, Boston College Economics and DIW Berlin
An Introduction to Modern Econometrics Using Stata:

* For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index