Browsing by Subject "synthetic control model"

This thesis asks whether Finland should experiment with unemployment insurance (UI) benefit frontloading. To answer this question, a comprehensive literature review on the matter is conducted. Covering both theoretical research and empirical frontloading studies, the subject matter is further examined by performing synthetic control method simulations on frontloading experiments in Sweden and Hungary. The data used in the empirical analysis extend from 1990 to 2020 and is hand-picked from freely available WorldBank databases. The question of frontloading has repeatedly came up in the Finnish political discussions during recent years. Experimenting with frontloading has found support among some government parties, but most concrete plans for reforming the UI benefit schedule have came from the opposition, namely from the National Coalition Party. The governing parties have struggled finding a concensus on the frontloading question, mostly due to lack of empirical evidence on its consequences. This thesis sheds light on what measures have been used when the Finnish UI system has been reformed earlier during the 21st century, and what response the Finns have presented for previous adjustments in the UI benefit system. It also discusses on the potential outcomes of the National Coalition Party’s proposals based on the theory, and highlights the effect on overall generousity of Finnish UI system to be negative. Simultaneously, this thesis presents that different frontloading method in Hungary had the desired employment effects while maintaining the overall generousity, resulting in a Pareto-improvement for the society. This thesis contributes to existing UI literature by not only connecting the theory of optimal UI and job search models to empirical experiments in Sweden and Hungary, but also by presenting synthetic control design methods to unemployment rate analysis. Empirical simulations in this thesis from two unconnected frontloading experiments reveal the special preparations needed to work with as volatile variables as the unemployment rate. Even though neither of the simulations resulted in statistically significant results, the methods presented display important features of usage of synthetic control design with unemployment rates concerning future studies.