Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: Seasonal Variables-Differencing or Dummies

From   "Nick Cox" <[email protected]>
To   <[email protected]>
Subject   st: RE: Seasonal Variables-Differencing or Dummies
Date   Thu, 5 May 2005 19:12:56 +0100

I am probably missing the point of this 
question, but I don't think you can list the circumstances 
in which S. is useful or valid. For example, 

use, clear
tsline S.air

I just used S. to help visualize the non-trend 
component, mostly seasonal. 

In short, S. is likely to be useful 
whenever you want it. No model need be in sight. 

There is a vast book on time series by Hipel
and McLeod with environmental applications. 
I'm pretty sure they discuss seasonality. 
Franses is another author who springs to mind. 
A search on Amazon would no doubt yield others. 

[email protected] 

> Please help me clarify my understanding in the correct procedure
> to use seasonal variables in time-series regressions. My very
> basic understanding of the correct use of the seasonal
> differencing technique (time series operator _S._ in Stata) is in
> the autoregressive integrated moving average (ARIMA) models. For
> other regression techniques in analyzing time-series, and panel
> data, the seasonal dummy variables are used.
> Hamilton (my much dogged-eared version for Stata 7) explains how
> to use _S._ in Stata, but not the ?why,? or the ?when? to use it.
> Harvey explains briefly both procedures and says that the
> differencing is used when the seasonal pattern changes over time.
> Thus, in this sense, the differencing would cause the seasonal
> pattern to become stationary.
> Is there ever an appropriate situation to use the _S._ time-series
> operator outside of an ARIMA model? In answering this specific
> Stata question, I ask that you extend your answer to an
> explanation, and possible reference that I could research, in
> controlling seasons (and cycles) in time-series data that goes
> beyond the usual econometric textbook treatment. An example, which
> does not necessarily need to be economic, would be very much
> appreciated.

*   For searches and help try:

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