|
||||||||||||||||||||||||||
|
8th OxMetricsTM User Conference 18 - 19 March 2010 George Washington University ,
Notice of meeting and CALL for papers
Contents Conference Overview This is the first announcement and call for papers for the 8th OxMetrics User Conference, to be held at
The conference will provide a forum for the presentation and exchange of research results and practical experiences within the fields of computational and financial econometrics, empirical economics, time-series and cross-section statistics and applied mathematics. The conference programme will feature keynote presentations, technical paper sessions, workshops, tutorials and panel discussions. We expect some of the OxMetrics’ developers (Jurgen A. Doornik, Sir David F. Hendry, Siem J. Koopman and Sébastien Laurent) to be present as keynote speakers. The other keynote speakers will be announced soon. The conference is open to all those interested, not just to OxMetrics™ users, from academic and non-academic organisations. There will be a session for participants to put their comments to the development team and suggest improvements. For details of past conferences, consult http://www.cass.city.ac.uk/faculty/g.urga/info1.html For updates on this conference, consult this page Presentation topics and abstract submissions
might include:
Submission of an abstract/paper. Please submit an abstract and full paper by email to Prof Frederick L. Joutz (bmark@gwu.edu) on or before 31 December 2009. The abstract should be no more than 200 words, and the paper should be in PDF format. Decisions on submissions will be made by 19th February 2010. You should indicate whether you wish to give:
Presenters of papers will have their conference fee reduced by 50% (one presenter per paper only).
The logistics of the conference is organised by Timberlake Consultants (www.timberlake.co.uk or www.timberlake-consultancy.com), the publishers of OxMetrics™. The contact for this meeting is Noelia Germino , Tel: +908 686 1251, Fax: +908 686 2307, Registration on line is available now. There will be a small conference fee: $140 if full payment of registration is received by 19 February 2010 or $160 if full payment of registration is received after 19 February 2010. This will cover costs for coffees, teas, luncheons and conference dinner on the 18th March 2010. A limited number of participants can stay at University accommodation and this can be arranged through Timberlake Consultants. The conference fee for full time students is $30. This fee does not include luncheons or conference dinner. However, if the student wants to have luncheon and conference dinner, their fee will be reduced by 25% ($105 and $120 respectively). The conference fee covers conference materials and coffees and teas. Presenters of papers will have their conference fee reduced by 50% (one presenter per paper only). OxMetrics™ is a modular software system providing an integrated solution for the econometric analysis of time series, forecasting, financial econometric modeling and for the statistical analysis of cross-section and panel data. The OxMetrics modules are: Ox Professional™ , PcGive™ , STAMP™, G@RCH™, SsfPack™ and TSP/OxMetrics™. OxMetrics is available within an user friendly GUI interface and extensive graphical reporting. Although OxMetrics does not finish there - it also offers a powerful module - Ox Professional, an object-oriented programming language - that allows power users to implement latest theoretical developments and/or integrated packages. OxMetrics Enterprise Edition™ is a single product that includes and integrates all the important components for theoretical and empirical research in econometrics, time series analysis and forecasting, applied economics and financial time series: Ox Professional, PcGive, STAMP, G@RCH. Purchasing the OxMetrics Enterprise Edition will provide the users with a very powerful and cost effective tool to use during their modeling work. In addition to the usual features in modern econometric software, OxMetrics Enterprise includes Autometrics™ - Automatic Model Selection (in PcGive). Autometrics™ Autometrics is a new algorithm for model selection within the general-to-specific framework, following on the work by Hoover and Perez (1999) and Hendry and Krolzig, 2005 See http://www.economics.ox.ac.uk/hendryconference/Papers/Doornik_DFHVol.pdf for further details. The OxMetrics™ software provides an integral solution for the econometric analysis of time series, forecasting, financial econometric modeling and for the statistical analysis of cross-section and panel data Social events Also, for those interested in doing so, there will be the option to attend a performance of Rachmaninoff's Vespers at 7:30pm on the same evening, following the conference dinner. The performance is by the Choral Arts Society of Washington, a well-known symphonic chorus in Washington DC (see http://www.choralarts.org/Season-Calendar.aspx); and one of the local organizers (Neil Ericsson) sings in the chorus. When you register for the conference, tick the special event box if you are interested in attending the concert. We are attempting to organize specially priced tickets for the concert, and we will let you know the prices of the tickets as soon as we finalize the negotiation. Empirical Model Discovery
I re-interpret model evaluation as discovering what is wrong, robust statistics as discovering which sub-sample is reliable, non-parametrics as discovering the functional form, and model selection as discovering which model best matches the criteria. It is crucial to tackle all of these jointly. Automatic methods enable formulation, selection, estimation and evaluation on a scale well beyond the powers of human intellect, including when there are more candidate variables than observations, and multiple structural breaks. I explain how major recent developments facilitate the discovery of models facing uncertainty over the choice of variables from a large candidate set, despite the high dimensionality, non-linearity, inertia, endogeneity, evolution, and abrupt change characteristic of economic data, all of which problems interact to make modelling so difficult in practice. Live computer illustrations using Autometrics show the remarkable statistical efficiency, power and feasibility of the approach, allowing economic-theory models to be embedded within general formulations, with post-selection bias corrections, and revealing small costs of search.
Robust Estimation of DCC - GARCH Models, The use of DCC models has now become standard in the financial econometrics literatures. Their estimation is usually done in two or three steps by Gaussian quasi-maximum likelihood. We show that this method is very sensitive to outliers in the data. We propose to use robust estimators for both models. The Monte Carlo study and empirical application document the good robustness properties of this estimation method. Empirical Model Selection: Friedman and Schwartz Revisited Using annual data from Friedman and Schwartz (1982), Hendry and Ericsson (1991) developed an empirical model of the demand for broad money in the United Kingdom over 1878--1970. Model selection was a central issue in assessing the merits of their model, so this paper re-evaluates that model with Autometrics, a recent third-generation algorithm for computer-automated model selection. Hendry and Ericsson's model is remarkably robust to the model selection path, as characterized through variations in the algorithm's settings for target size, pre-search, fixity of regressors, representation of the general model, and impulse saturation. This paper also assesses the empirical merits of Autometrics, using it to improve upon Hendry and Ericsson's (1991) model. |
