Growing up, Andriy Norets seldom heard talk about the economy of his native Ukraine in the media or elsewhere; it simply wasn’t part of the vernacular. But as the country transitioned to a market-based economic system and he began to hear about it more, Norets realized that he could turn his strength in mathematics into a career with practical applications.
“For me, economics provided this combination of still being able to do mathematics but at the same time applying it to real world problems — or in my case, providing useful methods for people that actually solve real world problems,” Norets said.
Following an undergraduate degree in applied mathematics, Norets earned his master’s in economics from the Kiev School of Economics, before going on to the University of Iowa for his Ph.D.
Today, Norets, who joins the Brown faculty as associate professor of economics this fall, is an econometrician with expertise in Bayesian econometrics. Several colleagues in the Department of Economics share his areas of interest.
“There is a tradition of Bayesian econometrics at Brown, and in general, econometrics is a very strong field here,” Norets said. “I’m very excited to be a part of this group now.” Norets comes to Brown from the University of Illinois at Urbana-Champaign.
In the spring, Norets will teach a graduate course on econometrics, which will focus on simulation methods, dynamic discrete choice models, and the Bayesian approach to econometrics, among other concepts.
One of those concepts — dynamic discrete choice models — was the focus of Norets’ early work, including his dissertation. Economists often use these models to describe how firms make a decision (enter the market or not? export or not?) or to model people’s behavior (should I retire this year or postpone it? should I take this job offer?) Using these models requires not only solving the problem but also estimating the model’s parameters, an exercise that can be quite difficult from a computational perspective. Norets’ work involved developing methods that combine the solution and parameter estimation into one efficient procedure.
Norets’ most recent work focuses on studying the data analysis methods used by economists and trying to figure out which procedures have properties that are useful from both the Bayesian and classical approach.
The latter, which is the approach most traditionally taught in classrooms, looks at how estimation procedures behave on average across different data sets. The former, an old approach that only recently became practical for a wide range of problems due to advances in computing speed and simulation methods, aims for a reasonable behavior for any given data set.
Norets developed estimation procedures that have properties of both the classical and Bayesian approach. Applied to the econometric problems that people often encounter, the results are not only reasonable but often work better than the approaches commonly used in economic research.
“Thinking about this Bayesian approach and this classical approach and how they relate to each other is a quite exciting area and there’s a lot of work to be done,” he said. “One of my objectives is to get people in economics to pay more attention to these methods, and I hope I can achieve that through my research.”