In our latest alumni interview, PhD in Economics program alumnus Pavel Čížek, an Associate Professor at the Department of Econometrics & Operations Research at Tilburg University, shares his journey from mathematical engineering to economics and explains the impact of semi-parametric and robust econometrics.
You completed an MA as well as a PhD in Mathematical Engineering. Why did you decide to pursue a PhD in Economics at CERGE-EI?
I indeed started my university studies in the field of applied mathematics within the program Mathematical Engineering at the Czech Technical University, and after the first two years, I focused more and more on probability and mathematical statistics. Although mathematical statistics falls under the label applied mathematics, this does not really imply that there has to be some applied work involved in the usual sense of the word – my studies were almost entirely focused on statistical theory.
Just as any other graduate at the end of university studies, I started to wonder what to do afterwards. This included intensive thinking about how and where to apply my statistical knowledge, and although there are many options including chemistry or medicine, economics was the field that got my interest due to its connection to and modelling of many personal decisions concerning education, employment, family, and so on. In Prague, the best place to study modern economics was CERGE-EI and its English PhD program looked very attractive to me. At the same time, CERGE-EI was also highly recommended to me by my supervisor prof. Jan Ámos Víšek, and so I had no doubts about it being my top choice.
Can you share your career path with us? Why did you decide for a career in academia?
My PhD studies proceeded in two phases. Although I spent the coursework years and the first year of thesis work at CERGE-EI in Prague, I later went on an exchange to Humboldt-Universität zu Berlin. After that, I spent there another four years as a teacher and researcher, working on the development of statistical software as well as on my research and PhD thesis. Given this experience working in academia, I tried to complement it with knowledge of other jobs and working environments I was interested in such as central banks or policy institutes. In this respect, my friends and colleagues from CERGE-EI were of great help by sharing their experiences obtained in their own careers. In the end, the freedom to pursue my own research and to be one’s own boss in this way was and still is most attractive to me, which strengthened my decision to pursue a career in academia.
Reflecting on your journey from starting your PhD in 1996 to your current position at Tilburg University, what are some key milestones or experiences that have greatly impacted your career?
From the very start, a relatively large PhD program with many students forming a vibrant community of students from diverse backgrounds provided a very rich and inspiring environment both for my personal development and my initial steps towards independent research. Next to my classmates, there were also many notable faculty members, but I was most influenced by meeting first as a teacher and later as a supervisor prof. Štěpán Jurajda. Our interactions deepened my interest in microeconometrics and also in pursuing an academic career, and his guidance was invaluable on this journey.
Later during my work at the Economics Faculty of Humboldt-Universität zu Berlin my collaboration with prof. Härdle and others at his chair (another word for (a part of) department headed by one full professor working with/leading team with lower rank academics; it is a common academic structure in Germany) greatly influenced me and introduced me to the world of non-parametric and semi-parametric econometrics and its applications in economics and finance. Although this field is not my sole research focus, it has been present in my research ever since.
The other two important steps in my academic life were already connected to Tilburg University. Both during the economic job market and just after getting tenure, I was deciding on where to pursue my academic career. In both cases, I decided to first start and then continue my work at the department of Econometrics and Operations Research at Tilburg University. Its relatively large group of faculty members oriented towards both applied and theoretical econometrics provided a similar rich research environment to what I was used to and appreciated during my early years at CERGE-EI.
Your research interests focus on semi-parametric and robust econometrics. Could you explain these methodologies and their significance in microeconomics and finance?
Most empirical research is linked to economic theories, which inform us about the links between key economic variables and thus provide us foundations for the model structure and interpretation. There are, however, many other characteristics of, for example, firms and individuals that play a role in financial and microeconomic decisions and lead to heterogeneity in empirical models. When these other characteristics are observed, they are typically used as controls, while the unobserved ones are described by some statistical assumptions. In this context, the non- and semi-parametric econometrics and robust econometrics aim at similar goals.
First, note that the non- and semi-parametric identification indicates which key variables and parameters of interest can be identified just with the structure derived from economic theory and with the minimal assumptions on the controls and unobservables. These results inform us about the knowledge that can be extracted from the empirical data without additional assumptions.
Next, both semi-parametric and robust econometrics provide estimation tools applicable under weaker statistical assumptions, although each is focused on somewhat different circumstances. In both cases, we recognize that real data do not follow models with ideal statistical assumptions and can deviate from them. While not common in highly aggregated macro-level data, such deviations can arise from heavy-tailed distributions of unobservables in financial data, from misreporting in individual-level survey data, from individual deviations from the behavior described by the model, or from other types of heterogeneity unexplained by the model.
On the one hand, robust econometrics addresses scenarios without sufficient information about these deviations, and given the lack of information to model them, designs methods that are insensitive to these deviations. For example, the behavior of individuals or firms during the recent COVID pandemic was affected by a wide variety of policy measures. In annual data, these effects might become just a few data points that deviate from the usual model structure and that cannot be easily modelled due to their small numbers. Robust econometrics then facilitates reliable estimation even in the presence of such data points that do not follow the specified model.
On the other hand, semi-parametric econometrics addresses scenarios where there is sufficient information about the above-mentioned model deviations (e.g., in large financial or marketing data, but also in institutional labor or health data). Semi-parametric econometrics then provides a methodology that facilitates estimation under minimal assumptions on the model structure and on the unobservable heterogeneity by directly estimating one or both of them using a large amount of information available in the data. This also includes recent development and applications of machine learning methods in microeconomics and finance.
In both cases, these methodologies let us focus on the relationships between important economic variables without imposing too strict or unnecessary statistical assumptions.
What do you perceive as the strongest message CERGE-EI gives to its students?
On a more general level, the economic questions, problems, and their solutions of our mainly economy-oriented society can be and should be answered, critically analyzed, and empirically judged using analytic and quantitative tools and observed data. Or in other words, while economic theories are important, they have to be grounded in and tested against the observed world using quantitative skills and methods. On a more personal level, the message includes the belief in the quality and rigor of the training at CERGE-EI, and consequently, the trust in our own ability to achieve our aims and goals, be it in a personal, national, or international context.
