Institutions are humanly devised constraints that shape interactions between people. Changing those constraints affects people's incentives and, therefore, affects economic, political, and social outcomes. Studying institutional arrangements helps to shed some light on why there is a high variation of the level of economic development across countries. These theses address the questions of how institutions are formed, how institutional changes affect incentives, and how they influence economic development.

The first chapter studies the effect of change in the rule that assign points in soccer on optimal strategies of soccer teams playing in a tournament. It demonstrates that the change in the rule increases incentives of teams to collude in order to trade points. It also has heterogeneous effects on top and lesser teams.

Second chapter looks at impact of good regional governance infrastructure on inflows of foreign direct investment (FDI) in 24 transition countries from 1993 to 2003. The model takes into account spatial spillovers and spatially correlated error terms. It is estimated by a recently developed generalized method of moment (GMM) three-stage procedure. The results show that the regional quality of institutions is an important factor that explains variations in FDI inflows. The positive effect of good regional governance dominates the effect of better developed regional markets.

The third chapter investigates determinants of the quality of governance inside a country. The main finding is the importance of relative geographical location: good governance in the neighboring countries has a positive impact on quality of governance inside a particular country. Spatial links work mostly through long-term determinants of governance that include culture, legal system, and colonial history. At the same time, the closest neighbors have the strongest impact on quality of governance, while cultural and colonial "neighbors" that are not close geographically, have smaller impact on the local institutional development. According to our results, cross-country regressions that do not take into account spatial interdependence of countries produce biased estimation of the coefficients and incorrect inference of variance-covariance matrix.