In this thesis I present three chapters that explore various themes in competition with a focus on collusion and spatial competition. Chapter one examines whether the availability of late-stage settlements for defendants in criminal price-fixing suits has a negative impact on the effectiveness of early stage leniency programs in the context of antitrust enforcement. Our main finding is that an appropriately designed settlement program can make collusion more difficult: in equilibrium, the adoption of an optimal settlement program by an antitrust authority (AA) reduces the occurrence of cartels by decreasing the long-run gains from collusion. However, overly generous settlement policies may undermine leniency programs and encourage the formation of more cartels Chapter two explores the relationship between business-cycle fluctuations and collusive behavior. From a theoretical perspective we demonstrate that the degree of antitrust enforcement (external cartel stability) directly influences the boundary that determines whether positive demand shocks are either pro-collusive or anti-collusive. We find that as cartels become increasingly unstable, they have a preference for defecting in the presence of positive demand shocks since there is riskiness associated with continuing to collude under a strong enforcement regime. From an empirical perspective, we find that the observed collusive activity is weakly procyclical, however, much of the variation is explained by enforcement and monitoring policies. Chapter three explores spatial competition in the Canadian banking industry at the 3-digit postal-code level. We study a two-way fixed-effects spatial panel model that disentangles market concentration (i.e., concentration ratios and the Herfindahl-Hirschman Index (HHI)) from the lesser explored notion of geographic concentration. This distinction is relevant since bank branches tend to operate in spatial clusters that form sub-geographic markets, thus, altering competitive outcomes. We find that the HHI and spatial clustering are both negatively correlated with the relative size of the bank-branch network. In effect, branch proliferation is strong in geographic regions that are more competitive and more uniformly dispersed spatially.