12) University of Lodz

Scientific Staff

Prof. Aleksander Welfe
Wojciech Grabowski
Piotr Keblowski
Anna Staszewska

Aleksander Welfe graduated in economics (econometrics) from the Warsaw University. He passed his habilitation in economics and econometrics at the Wroclaw Academy of Economics. Currently, he is ordinary professor of economics at the University of Lodz and the Warsaw School of Economics.

Expertise of the Team

The team members are all employed in the Chair of Econometric Models and Forecasts, which has established a very strong position amongst the Polish econometric research centres for its research in macroeconometrics and the econometrics of financial markets. The team uses the most recently developed techniques for time series analysis, the scope of the models varying from large-scale macromodels to partial models of market and financial processes. The ongoing task of the team members is to improve the inference techniques used. The objective has been either to propose new inference procedures or to study and improve the small-sample properties of established techniques. The methods used include the computer-intensive techniques of Monte Carlo experimentation, bootstrapping, EM and ECoM algorithms and also genetic algorithm and other optimization heuristics. Most recently the focus of the team members has been on improvements to integration and cointegration inference (Keblowski and Welfe (2004), Keblowski (2005)), modelling of nonstationary censored data (Welfe and Grabowski (2006)) and response analysis for multivariate dynamic models (Mizon and Staszewska (2006), Staszewska (2006)).

The Chair of Econometric Models and Forecasts organises two annual conferences: the international Macromodels conference (which now has a history going back 32 years) and the PhD Workshop - a national conference for young econometricians and statisticians.


  • Keblowski P.: “Small Sample Power of Bartlett Corrected LR Test of Cointegration Rank”, Unpublished Manuscript, http://www.ualg.pt/feua/conf/urct/prog/ps1/p1027.pdf, 2005.
  • Keblowski P. and A. Welfe: “The ADF-KPSS Test of the Joint Confirmation Hypothesis of Unit Autoregressive Root”, Economics Letters, 2004.
  • Mizon G. E. and A. Staszewska: “Conditional Response Analysis”, in G. Phillips (ed.) “The Refinement of Econometric Estimation and Test Procedures: Finite Sample and Asymptotic Analysis”, Cambridge University Press, forthcoming 2006.

Expertise in Training Young Scientists

Aleksander Welfe supervised about 20 students in Ph.D. courses. During the last year, 6 of them prepared their theses, successfully passed the exams and were awarded Ph.Ds.

Links within the Network

There exist several informal links between the team members and other network participants based on meetings at conferences, e.g., the annual Macromodels conference where the work of the team members as well as that of Winker, Gilli, Maringer and Meyer has been presented. In the past, Welfe was a faculty member of the University of Konstanz supported by an Alexander von Humboldt grant and was closely co-operating with Winker.

Role of the Research Team

The team members will build on their existing work in the fields of cointegration analysis, modelling of non-stationary censored data and response analysis. They will apply optimization heuristics to these areas. In particular, the team members will consider the construction of confidence bands for impulse response functions by means of optimization heuristics. Furthermore, the team will work on model selection issues for modelling macroeconomic data with long memory (cointegration, fractional cointegration). A specific focus will be on performance for short time series, which are typical of contemporary applications for new member states. For this purpose, extensive Monte Carlo simulations and a proper presentation of the results will be required. Finally, the issue of non-stationary censored data will be approached which also results in highly complex optimization problems which cannot be tackled by standard methods. Both applications for estimating the parameters of this type of models and for inference will be developed.