News and Activities

News

• (03/10/2007) Release of STAMP 8 in October 2007. List of new features can be found here. More updates will be available soon.

• (19/04/2007) The data used in Stamp 7 are updated and are available for download.

• (02/10/2006) Martyn Duffy has used Stamp in his paper `Tobacco consumption and policy in the United Kingdom'. This paper is published in Applied Economics, 2006, 38, 1235-1257.

• (10/07/2006) Screenshots of Stamp 7 are added.

• (22/06/2006) The Stamp 7 manual is available now. Visit the website of Timberlake Consultants for more information.

• (22/06/2006) Stamp 7 is released in April 2006.

• (22/06/2006) 26-30 June 2006: OxMetrics events in Japan. Visit the website (in japanese).
  • 26 June: Bank of Japan
  • 27 June: Hitotsubashi University
  • 28 June: Japan Economic Research Center
  • 29 June: Central Research Institute of Electric Power Industry. This event can be attended by anyone who is interested in OxMetrics. Please visit the website for more information.
  • 30 June: Economic and Social Research Institute, Cabinet Office, Government of Japan

• (22/06/2006) 14-15 September 2006: 4th OxMetrics Users' Conference, Cass Business School, London. Call for papers coming soon. Information for the previous conference is still on the 3rd conference web site.

• Software Review: Structural time series modelling with STAMP 6.02, Journal of Applied Econometrics, Volume 20, Issue 4, pages 571-577 (2005) by Gilles Teyssière.

• Readings in Unobserved Components Models edited by Andrew Harvey and Tommaso Proietti has appeared. This book presents a collection of readings which give the reader an idea of the nature and scope of unobserved components models and the methods used to deal with them.

New features in Stamp 7

• Time-varying regression coefficients can be included as part of the model.Three specifications can be selected: (i) random walk, (ii) spline sprecification (or smooth trend) and (iii) return-to-normality (deviations from a fixed coefficient follow an autoregressive process of order 1).

• Missing values can be treated in this version of STAMP. In all levels of the analysis, missing observations are accounted for automatically. Model-based estimates of the missing values can be produced. Confidence intervals can be included in the graphs with estimated components. The predicted, filtered and smoothed estimates of the components can be presented simultaneously.

• The observation weight functions that are used for the estimation of the unobserved components are given as output. The associating spectral gain and phase functions can also be produced.

• Model-based forecasting and backcasting can be carried out.

• Unobserved stationary autoregressive processes of orders 1 and 2 can be included in the model together with three stochastic cycle components.

• Higher-order smooth trends and cycles with band-pass filter properties can be included in the model.

• The trigonometric seasonal component consists of separate processes for the different seasonal frequencies. These processes can be selected separately in the model. Also different seasonal variances can be attached to the seasonal processes.

• More flexible options for the handling and estimation of parameters are provided. The parameters can be treated without transformation.

• Dates for outliers and breaks can be stored and remain part of the model once selected.