Master of Science : 1990, econometrics, statistics, and demography, SGIPS (Warsaw School of Economics), Ph.D. : 1995, University of Warsaw, Faculty of Economic Sciences, thesis title: “Growth-Oriented Structural Adjustment Policies for a Dual Economy. An Econometrical Assessment”.
Associate professor : 2020, University of Lodz, Poland
Habilitation degree research field : « Non-Extensive Entropy Econometrics for Low Frequency Series: National Accounts-Based Inverse Problems ». This proposed subdiscipline of econometrics generalizes the maximum entropy econometrics on nonlinear systems using Tsallis entropy and statistical information theory formalism.
From 2016 :
Consultant – statistical analysis and methodology,
Polish Statistical Office
International expert at the Polish Accreditation Committee (minister of higher education), Warsaw.
The main goal of our scientific and research activity is reflected in the publications presented by us in recent years. Familiarization with the important concepts of nonadditive thermodynamic statistics has directed us interest to study non-ergodic (inverse-problems) phenomena characterized by greater complexity than those described in the attractor of the Lindenberg central theorem-limit, of which Shannon’s entropy and most traditional econometric models are limited to. In the centre of non-extensive statistics and entropy of Tsallis points out a PL distribution. After learning about the properties of this entropy, about its more general character in comparison with Shannon’s entropy, we came to the conclusion that its connection with Kullback-Leibler-Jaynes statistical information formalism should provide an econometric modelling tool potentially capable of more general analysis of systems including non-ergodic ones, examples of which are fractal and multifractal phenomena. These premises have conditioned the main scientific goal of our research over the past years.
Thus, the rationale of proposing the methodology of non-extensive (cross-) entropy econometrics (NCEE) remains the fact that in the real world, socioeconomic rare events may have higher impact than more frequent events could when seen with respect to Gaussian distribution. As alluded to in our recent publications, long-range correlation and observed time invariant scale structure of high frequency series may still be conserved—in some classes of nonlinear models—through a process of time (or space) aggregation of statistical data. Our research shows that such a process is better described by a power-law (PL)-related model belonging to Tsallis non-additive statistics. The methodology extends PL to the Kullback-Leibler-Jaynes statistical theory of information. Next, this formalism allows for connecting to the system an econometric model in the Bayesian context. In this context, an ad hoc new statistical inference, different from the Neyman classical one based on the Lindeberg central theorem-limit, has been proposed to allow for real world implementation of the methodology. Because of this and based on existing literature, the approach is termed Non-extensive (Cross-) Entropy Econometrics or using a less technical expression, Superstar-Generalized Econometrics, to emphasize the role in a society( or elsewhere) of “superstars” (sparse or rare events, e.g., high earnings in financial markets, the economic consequences of a hurricane, earnings of artists or CEOs of large firms with respect to exceptional popularity, etc..). Traditional econometrics generally considers that outlier data should be excluded from the original data before any econometric modelling.
Our current research activity extends the recently proposed model to simultaneous equation models, based on the presumption in real life of interconnected non-linear phenomena, at different levels of complexity.