1.    The first article of the doctoral thesis is titled “Forecasting the Estonian rate of inflation using factor models”. This paper investigates the forecasting performance of time series models in which large amounts of information are summarized in a smaller set of factors. The factors are incorporated in a vector autoregressive model in order to forecast the Estonian headline and core inflation rate during a period of low and stable inflation. The results show that forecasting the Estonian rate of inflation remains challenging. While the forecasts from a factor-augmented vector autoregressive (FAVAR) models can improve upon the forecasts of simple univariate autoregressive models in some specifications, consistent improvements cannot be achieved across all potential model specifications. Second, the paper shows that even though the general FAVAR forecasting equation is based on a theory-agnostic approach, defining homogenous subsets of variables can be used to create sets of sharpened factors, which can considerably improve the forecast performance of the FAVAR models. 
2.    The second publication included in the thesis is titled “Noise shocks and business cycle fluctuations in three major European economies”. This publication studies how supply and demand noise shocks contribute to business cycle fluctuations in France, Germany and Italy. Noise shocks are interpreted as misperceptions about supply factors such as technology and demand factors such as consumption preferences or prices. These misperceptions can drive the dynamics of GDP growth and inflation even absent changes in the underlying fundamentals of the economy. The empirical estimation strategy relies on a sign-identified vector autoregressive model, where the restrictions are derived from a theoretical macroeconomic model. The main results show that GDP growth reacts positively both to a supply noise and a demand noise shock in German and France while the results for the response of GDP growth to a demand noise shock are statistically insignificant for Italy. The expansionary effect of GDP growth following demand noise shocks is at odds with evidence from studies focusing on the US but can potentially be rationalized with differences in how monetary policy is conducted or how the information set of firms in European economies are formed. The publications also shows that around one third of the fluctuations of GDP growth in the three countries are driven by noise shocks, while fundamental shocks are responsible for the remaining two third of the output fluctuations. 
3.    Publication three is titled “Alternative frameworks for measuring credit gaps and setting countercyclical capital buffers” and explores whether alternative measures of credit-to-GDP gaps can be used to inform the policy maker about the stance of the financial cycle. While the Basel 3 credit-to-GDP gap is the got-to metric for the policy maker to assess the position of the financial cycle, it suffers from a range of shortcomings. Publication three addresses some of those shortcomings and shows that measures which do not rely on some sort of a statistical filter can at least track the Basel 3 credit-to-GDP gap in a large sample of European economies. For countries in Eastern Europe with shorter time series, the alternative measures often show more desirable properties in terms of ex-post signaling performance than the Basel 3 credit-to-GDP gap. Finally, the publication shows that if capital buffers would have been implemented based on the estimation of the credit-gaps from the alternative measures, the buffer sizes would behave similarly for many countries in the sample.  
4.    The fourth publication in the thesis is titled “Banking Sector Concentration, Competition and Financial Stability: The Case of the Baltic Countries” and analyses the relationship between different measures of banking sector concentration and different measures of banking sector risk in Estonia, Latvia and Lithuania. This topic is of major importance as the three Baltic countries have extremely high concentrated banking sectors. The publication provides empirical evidence of a non-linear relationship between banking sector competition and banking sector stability. This means that both high and low levels of banking sector competition are associated with a higher risk of bank-level insolvency. The results also show that the proxies for banking sector competition and banking sector concentration do not necessarily capture the same dynamics of the banking market structure.  
5.    The fifth and final publication titled “The evolution and heterogeneity of credit procyclicality in Central and Eastern Europe”. This publication combines aspect of financial research with evidence provided in publication four and analysis of how the structure of the banking sector may affect the observed procyclicality of credit. The paper uses a panel vector autoregressive model to show that credit procyclicality was higher before the global financial crisis than after the global financial crisis. In the next step, 11 Central and Eastern European Countries are analyzed separately. The results show that a considerable degree of heterogeneity in credit procyclicality exists across the countries. In order to gauge the impact that the banking sector structure might have on credit dynamics, the final step of the analysis interacts the panel vector autoregressive model with a measure of country-level banking sector concentration. The results show that credit procyclicality is higher in less concentrated banking markets. 

Supervisors: 
Professor Karsten Staehr (primary supervisor), Associate Professor of University of Tartu Lenno Uusküla (co-supervisor) 
 
Opponents:  
Professor Jesus Crespo Cuaresma (Department of Economics, Vienna University of Economics and Business, Austria)
 
Assistant Professor Povilas Lastauskas, (Department of Business Analytics and Applied Economics, Queen Mary University of London, United Kingdom)
 
 
The thesis is available here: LINK
 https://digikogu.taltech.ee/et/Item/434dfb35-3c2b-4952-b432-ca5434ff13e4 

This work was supported by funding from the ASTRA “TTÜ arenguprogramm aas- tateks 2016-2022” Doctoral School in Economics and Innovation Project under Grant Agreement No. 2014-2020.4.01.16-0032.