Assistant Professor of Economics Artūras was awarded a Veni grant  for his project, “ We do not live in a bubble: economic shocks in misspecified panel data models ”. The personal Veni grants are worth up to a maximum of €250,000 and enable talented researchers who have just completed a PhD to conduct research of their own choice. They are part of the Innovational Research Incentives Scheme run by the Netherlands Organisation for Scientific Research (NWO).

Q. Tell us about your research.

A. We live in an open and interconnected environment, where no “bubble” protects us from constant changes. For example, the key country level economic variables (e.g. interest rates) are subject to global economic shocks (trends), thus countries cannot be studied in isolation. Common economic shocks also affect decisions made by individuals, e.g. occupational choices. As these shocks cannot be measured perfectly, appropriate procedures accounting for them have to be used in empirical analysis.

Fortunately, given that countries and individuals can be observed over multiple years, it is natural to analyse data of this type as double index (panel). The unobserved characteristics are then captured by using unit/time specific parameters. However, as the number of parameters increases with the sample size, the statistical analysis becomes challenging.

The literature on this topic has seen a major development in the last decade, especially after the Great Recession. But most of the procedures are developed assuming correct empirical specification, which is overly simplistic for complex economic data. In this project, I study how costly the relaxation of this assumption is, i.e. what kind of conclusions can be drawn in misspecified models? And are they of any relevance for policy makers and the society at large?

The theoretical arguments leading to the answers are established in several steps. At first, I investigate the sensitivity of already available procedures when some of the underlying assumptions are violated. Later, I suggest a set of new statistical approaches that are less sensitive to misspecification.

Q. What’s the biggest challenge of this research?

A. The biggest challenge of my project comes from the (statistical) theoretical aspects of this research topic. The current state of the art mathematical toolkit used by panel data researchers is not rich enough to cover all of the situations I want to study. More importantly, at this stage of the project it is unclear whether necessary (or sufficient) conditions are likely to be satisfied in data structures micro and macro economists are interested in.