I'm concerned to seasonal adjustment procedure and want to know the criteria for this purpose can anyone please give me the answer of the following question. what should be the criteria for seasonal adjustment whether to do or not to do?

  • 1
    $\begingroup$ You will need to give a lot more detail on what you are trying to do and what kind of seasonal adjustments you are considering. $\endgroup$ – mike1886 Jun 2 '14 at 15:46
  • $\begingroup$ @mike1886 i just want to know a rough criteria for seasonal adjustment,i am just a beginner :) $\endgroup$ – Rana Muhammad kashif Jun 2 '14 at 17:41

You are concerned that you have seasonality in you data. There exists a test-statistic called qs, that tests if there is any seasonality in your data or not.

The statistic is defined as follows:

$$ QS = n(n+2)\left( \frac{R_s^2}{n-s} + \frac{R_{2s}^2}{n-2s} \right) \,, $$

where $R_s$ and $R_{2s}$ denote autocorrelations of your series.


Seasonal adjustment is the process of removing a nuisance periodic component. If you have a time series with small deviations in it, the true trend can be masked by the seasonal component. Therefore, it may be beneficial to remove this seasonal trend. A nice overview of this can be found in the matlab tutorial: http://www.mathworks.com/help/econ/seasonal-adjustment-1.html

Without any further information on what you are modeling it is impossible to tell you if seasonal effects will matter. For example, assume your time series has a fluctuation of 1000 units, now assume the seasonal effect only accounts for 0.01% of this. At that point it probably is not important to model and remove the season component. On the other hand, if the seasonal effect accounts for 80% of the fluctuation then you do need to model and remove it.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.