DEMETRA +
by Eurostat and National Bank of
Belgium
In case of any problems or
questions, please contact
Eurostat Unit B2
Methodology and Research at: estat-methodology@ec.europa.eu
To download
documentation about Demetra+ from CROS portal please click here
Reference page to ESS
guidelines on Seasonal adjustment here
For info about the next ESTP
course on DEMETRA+ (for advanced users) that will be
held at Eurostat from 15 to 17 of November 2011 please consult here
Description
Seasonal adjustment is an important
step of the official statistics business architecture and harmonisation of
practices has proved to be key element of quality of the output. In this
spirit, since the 90s, Eurostat has been playing a role in the promotion,
development and maintenance of a software solution (Demetra) freely available
for seasonal adjustment in line with established best practices.
In 2008, ESS (European Statistical System) guidelines on SA have been
endorsed by the CMFB and the SPC as a framework for seasonal adjustment of
PEEIs and other ESS and ESCB economic indicators. ESS guidelines cover all the
key steps of the seasonal and calendar adjustment process and represent an
important step towards the harmonisation of seasonal and calendar adjustment
practices within the ESS and in Eurostat. A common policy for the seasonal and
calendar adjustment of all infra-annual statistics will improve the quality and
comparability of the national data as well as enhance the overall quality of
European to the extent that proper SA tools exist and are available.
The SA Steering Group (the Eurostat-ECB high level group of experts from
NSIs and NCBs which has produced the ESS Guidelines for seasonal adjustment) is
promoting the development of a flexible software solution for SA to be used
within the ESS. The group has drawn its attention on the object oriented
technologies used by the R&D Unit of the Department of Statistics of the
National Bank of Belgium to develop a series of prototype tools for SA. This
has been considered as an adequate framework for the cooperative development of
a new generation of sustainable SA tools, enabling the implementation of the
ESS guidelines and replacing the previous Demetra whose maintenance and
sustainability is put in question.
Demetra+ is a family of modules on seasonal adjustment, which are based
on the two leading algorithms in that domain (TRAMO&SEATS@ / X-12-ARIMA).
TRAMO&SEATS@ (TRAMO \"Time series Regression with ARIMA noise, Missing
values and Outliers\", and SEATS, \"Signal Extraction in ARIMA Time
Series\", developed by Agustín Maravall
and Victor Gómez) and X-12-ARIMA (developed by
David Findley and Brian Monsell) are two different
methods to seasonally adjust a time series. Both methods can be divided into
two main parts: a pre-adjustment step, which removes the
\"deterministic\" component of the series by means of a regression
model with Arima noises and the decomposition part itself. The two methods use
a very similar approach in the first part of the processing but they differ
completely in the decomposition part. Their comparison is often difficult, even
for the modelling step. More especially, their diagnostics focus on different
aspects and their outputs take completely different forms. One of the main
features of Demetra+ is to normalize - as much as possible - the different
methods. It tries to improve the comparability of the two methods by using as
much as possible, a common set of diagnostics and of presentation tools. That fundamental
choice implies that a number of routines of both methods have been re-written
in Demetra+. That can lead, compared to the original
programs, to small discrepancies in diagnostics or in peripheral information
that should not alter the general \"message\" provided by the
algorithms. Under no circumstances should the main results of the
original programs (seasonally adjusted series...) be impacted by that solution.
Features
The technology (Object Oriented components)
underlying the toolkit has proved to be a powerful and flexible solution for
managing the complexity of seasonal adjustment algorithms and integrating the
major well-known SA engines provided by the Bank of Spain and USCB. In
addition, it could easily be embedded in many different environments allowing
fast developments and extensions.
Future plans
JAVA version