There is a wide range of seasonal adjustment software and interfaces available to use. The summaries provided give links to where you can download the relevant software.
For official statistics, the two most commonly used seasonal adjustment methods are X-12-ARIMA and TRAMO-SEATS.
|Responsible||United States Bureau of the Census|
X-12-ARIMA is based on the X-11 filter based approach to seasonal adjustment. It uses linear filters to estimate and remove the systematic calendar related effects. It also combines the use of ARIMA models for forecasting and regression. It is flexible but also a very robust seasonal adjustment package.
There are many different platforms available: A native DOS executable, a windows interface, linux, and an implementation into SAS. Note that the different implementations may not have all the full options as used in the original programs direct from the United States Bureau of the Census website.
The manual included in the download is very comprehensive, both on the theory but also the options in the program, and is the best place to start if you're new to X-12-ARIMA.
X-12-ARIMA, along with TRAMO-SEATS, are jointly recommended seasonal adjustment packages for official statistics by Eurostat.
X-12-ARIMA is free to download and use.
|Responsible||Bank of Spain|
TRAMO stands for Time series Regression with ARIMA noise, Missing values and Outliers; and SEATS stands for Signal Extraction in ARIMA Time Series. The SEATS module performs the seasonal adjustment decomposition using a model-based approach. This is done by using filters which have been derived from ARIMA models describing each different time series component.
There is a DOS version and Windows version available. The official website gives some examples of use and we'd recommend checking the examples out first as some of the parameter settings are not the easiest to set optimally for the non-expert.
The program can handle seasonal adjustment for a large number of time series in a batch approach.
TRAMO-SEATS, along with X-12-ARIMA, are jointly recommended seasonal adjustment packages for official statistics by Eurostat.
TRAMO-SEATS is free to download and use.
|Responsible||Eurostat, user community|
This is free open source software. It can be used as an interface to both X-12-ARIMA and TRAMO-SEATS.
One of the main features of Demetra+ is to normalise the two different methods and provide ease of use for national statistics institutes. For example, by using a common set of diagnostics and presentation. In the implemention of Demetra+, both X-12-ARIMA and TRAMO-SEATS have been rewritten so the outputs may be slightly different from the native versions. There is a planned Java implemention of Demetra+ scheduled for release at the end of 2012.
|Responsible||Maintained by a group of academics|
|Opinions||STAMP stands for Structural Time Series Analyser, Modeller and Predictor. The estimation of the time series components and signal extraction is done using state space modelling methods and Kalman filtering for the estimation. It can be used through the Oxmetrics software although you'll have to pay for this.|
|Opinions||This is free open source software. It can be used as an interface to both X-12-ARIMA and TRAMO-SEATS. It can also interface to other statistical packages.|
|Software||proc x12 (within SAS)|
|Opinions||The SAS procedures are: proc x11 and proc x12, which are available within the commercial SAS package. Note that the results will differ from the native X-12-ARIMA version as they have been recoded within SAS. Many National Statistics Institues, for example, Statistics Canada, use SAS heavily in their production of seasonally adjusted estimates. You'll need to pay for SAS.|
|Opinions||Can perform seasonal adjustment with: X-11, X-12-ARIMA and TRAMO-SEATS. This is commercial software and you'll need to pay for it.|