The law of agency is one of the basic building blocks of business. Agents facilitate transactions between firms and structure governance relations within firms. Agency law regulates both types of activity. Yet agency law's main focus has always been on the transactional side. The key feature of transactional--traditional--agency law is the tripartitite relationship among the principal, agent, and third party. Law and economics scholars, however, have mostly focused on the role of agency within organizations. More traditional business law scholars have also paid little attention to agency law in recent years. This paper develops an economic theory of transactional agency law. I argue that what makes the three-party structure of agency law different from two-party structures is the potential for collusive behavior between any two of the three parties against the other. My thesis is that transactional agency law can largely be understood as attempting to facilitate transactions by deterring collusion between principals and agents, between agents and third parties, and between principals and third parties. Agency law does this by providing guidelines for determining when collusion of various types should be presumed. I develop a theory of collusive behavior and then apply this theory to various doctrines in agency law that have long troubled scholars and courts, including contractual authority, undisclosed principal, and ratification. Throughout, I compare the collusion theory with an alternative economic theory that explains agency rules under the least cost avoider principle.