Improving quality in seasonal adjustment in Short-Term Statistics using JDemetra+ regressors and TEAM R-package

Félix Aparicio Pérez; José Fernando Arranz; María Cruz Gómez Izquierdo; María Novás Filgueira; Juan Rebé Sancho; Elena Rosa Pérez; Carles Sáez Calvo; Luis Sanguiao; Teresa Vázquez

Ayuda

Improving quality in seasonal adjustment in Short-Term Statistics using JDemetra+ regressors and TEAM R-package

Felix Aparicio Pérez ^[1] ; José Fernando Arranz ^[1] ; María Cruz Gómez Izquierdo ^[1] ; María Novás ^[1] ; Juan Rebé Sancho ^[1] ; Elena Rosa Pérez ^[1] ; Carlos Sáez Calvo ^[1] ; Luis Sanguiao Sande ^[1] ; Teresa Vázquez ^[1]
1. [1] Time Series Analysis at Instituto Nacional de Estadistica (INE)
Localización: BEIO, Boletín de Estadística e Investigación Operativa, ISSN 1889-3805, Vol. 41, Nº. 2, 2025, págs. 39-46
Idioma: inglés
Enlaces
- Texto completo
Resumen
- Short-term business statistics (STS) are the earliest statistics released to show emerging trends in the European economy.
  
  Monthly and quarterly STS provide data for the main economic sectors: industry, construction, trade and services, excluding financial and public services.
  
  STS Regulation requires data that are calendar adjusted and calendar and seasonally adjusted, in addition to unadjusted data. Seasonal adjustment (SA) procedures eliminate the estimated seasonal and calendar effects from the original time series and obtain SA-estimates that are likely to reveal what is new in a time series.
  
  JDemetra+ is the seasonal adjustment software officially used in the European Statistical System. Among other methods, it allows the use of a model-based TRAMO-SEATS approach for performing seasonal adjustment of time series. In this approach, a RegARIMA model is fitted to the series. It also offers the chance to calculate regression variables to model calendar effects, including trading day regressors that take into account the composition of the days of the month.
  
  ARIMA models used to adjust STS time series play a very important role to obtain accurate adjusted data. But sometimes the work of updating the ARIMA models can become a burdensome task as the manual identification of a suitable model can become complex and time-consuming and automatic procedures can provide a model without taking into account some restrictions considered as essential for the domain expert.
  
  Time-Series Exhaustive Automatic Modelling (TEAM) is an R package developed by Statistics Spain, and based in the JDemetra+ ecosystem, that can help in the process of updating ARIMA models by providing a list of models ordered according to a global score.
  
  Methodologically, we can set a priori specifications (outliers, calendar regressors, maximum/minumum values of the ARIMA parameters), and the local scores by hierarchical levels can help us guarantee the quality. In this sense, one of the main advantages of TEAM is that the ARIMA models provided can be subject to some restrictions specified by the domain expert.
  
  To carry out calendar adjustment at Statistics Spain we have been using customized working days regressors. However, for time series of specific activities the residual effects of trading days were not being completely removed. Several analyses have been undertaken and improvements have been achieved when using JDemetra+ regressors.
  
  Due to the increase in quality, the JDemetra+ regressors and the use of the TEAM package to update the ARIMA models are included in the seasonal adjustment process as from January 2024 in STS at Statistics Spain.
Referencias bibliográficas
- Aparicio Perez, F., Arranz Arauzo, J. F., S ´ aez Calvo, C., Sanguiao Sande, L., & V ´ azquez Guti ´ errez, M. T. (2024, June 5-7). Time-series...
- Eurostat. (2015). Ess guidelines on seasonal adjustment (Tech. Rep.). Author. (Available at: https://ec.europa.eu/eurostat/ documents/3859598/6830795/KS-GQ-15-001-EN-N.pdf)
- Eurostat. (2017). European statistics code of practice (Tech. Rep.). Author. (Available at: https://ec.europa.eu/eurostat/ documents/4031688/8971242/KS-02-18-142-EN-N.pdf/e7f85f07-91db-4312-8118-f729c75878c7)
- Eurostat. (2018). Handbook on seasonal adjustment (Tech. Rep.). Author. (Available at: https://ec.europa.eu/eurostat/ documents/3859598/8939616/KS-GQ-18-001-EN-N.pdf)
- Eurostat. (2020). Commission implementing regulation (eu) 2020/1197 laying down technical specifications and arrangements pursuant to regulation...
- INE Spain. (2024). Ine standard on seasonal and calendar adjustment. (Available at: https://ine.es/en/clasifi/estandar efectos estacionales...
- National Bank of Belgium. (n.d.). Jdemetra+ reference manual [Computer software manual]. (Available at: https://jdemetra -new-documentation.netlify.app/)