seasonality after seasonal adjustment

[kanonathena]

Hi

After making seasonal adjustment to a series with ARIMA (no seasonal dummies or constant), the resulting series still has a seasonal patten as detected by regressing the adjusted series on seasonal dummies. why is this?

[trubador]

Seasonal adjustment is a data-dependent analysis, so it is quite difficult to understand the problem without seeing the actual data. Output reports of formal seasonal adjustment procedures (X12 or TRAMO/SEATS) are more conclusive and reliable. These reports provide very detailed information on the seasonal behaviour in your data (e.g. whether the data have a moving seasonality pattern). Seasonal dummies are simple dummy variables assigned to specific periods and may also capture movements other than seasonality. Since built-in seasonal adjustment procedures can also auto detect and remove any outlier effect, you should also check for any significant irregularities in your regression.

[bvguizar]

Hi,

I have made seasonality adjustment to my series with tramo/seats but keeps showing seasonality. Do you know how can I correct that?

Thank you

[trubador]

You can add seasonal dummies to your analysis.

[punjaphp]

How do I add seasonal dummies in eviews program?
thank you

[trubador]

For monthly data, you can start by @expand(@month) command and then keep the significant ones with @seas() command:

Code: Select all

equation myeq.ls y x @expand(@month)
equation myeq.ls y x @seas(1) @seas(5) @seas(8) @seas(10)

[punjaphp]

Great! thank you trubador

[bvguizar]

Hi,

What do we do after we run the OLS with the dummies. I found that all of the dummies are significant, then I go to X12 arima or how do I correct this? Please help