The present dissertation consists of three essays on the economics of collusion. The first essay examines bidding in U.S. Forest Service first price timber auctions in the Northern Region, where the potential for collusive bidding has been recognized. I modify the empirical methods in Porter and Zona (1993) and find a group of potentially cooperative (PC) bidders, who can submit complementary bids. Benefiting from a striking feature of data, where sealed bid auctions and ascending bid auctions were used side by side, I find further corroboration for my findings by analyzing PC bidders bids in ascending bid auctions.The second essay, joint with Shigeki Isogai, follows the empirical regularity noted by Marshall and Marx (2015). We present a reputation model in which a long-lived multi-product firm that is sequentially engaged in explicit collusion with short-lived single product firms can exploit the cartel leniency policy offered byantitrust enforcement authorities. The long-lived firm may have incentive to seek leniency to build and protect its reputation as a tough firm, who never tolerates any deviant conduct. This may help the long-lived firm deter deviations in the cartels. Our model provides a new insight on cartel firms incentive to report their own cartel, a potential counterproductive effect of the leniency policy, and important policy implications to the design of the amnesty program.The third essay analyzes the effect of an antitrust leniency program on the decision to merge or, alternatively, explicitly collude. Buyers use procurements but the procurement will be re-conducted when the buyer is dissatisfied with the bids of the incumbent sellers or a cartel is discovered. Additionally, production costs ofthe sellers may change in each round of the procurement. If the production cost states in the first round are unprofitable, the sellers may switch to the re-conducted procurement by reporting the existence of the cartel and committing to act noncollusively through leniency applications. Thus, sellers that were indifferent between merging and forming a cartel with no leniency program may prefer colluding in the first round procurement in the presence of a leniency program. So a leniency program may induce both more discovery of cartels as well as more cartel formation.