how to perform seasonal adjustment with tssmooth？

[Fuyang Zhao]

Hi everyone,
I want to seasonaly adjust a variable with tssmooth.

I typed

tssmooth s gdp_season=gdp,per(4)
  tsline gdp_season gdp

And the result is shown is the following figure. The adjusted data is almost the same as the original one. How could I do?

Graph.eps

The data is in the attachment. Can anyone help me?

[Nicolás Ronderos]

http://davegiles.blogspot.com.co/201...om-quandl.html

[Fuyang Zhao]

http://davegiles.blogspot.com.co/201...om-quandl.html

somehow i cannot open the link……

[Carole J. Wilson]

This looks like spam since the link says something about flowers for mom when I hover over it.

[sladmin]

We can confirm that the link above is not spam and is statistical (seasonal price adjustments for flowers around Mother's Day) in nature.

[Nicolás Ronderos]

Is not spam. The general idea of the S.A. procedure using tssmooth is to obtain the trend of the time series, then substract that trend from the original series (if it is assumed an addtive component, the best is to take logs of the series to make sure of this), then with the cycle component obtain seasonal factors by averaging the cycle component for each month, quarter, etc, then just substract the seasonal factors to the original time series.

The trend component is usually obtained by a moving average of order 11, and to force the seasonal factors to have zero mean you can estimate the mean of the seasonal factors and then just substract the mean from them.

This is the basic idea of any seasonal adjustment procedure.

Best regards,