ABSTRACT
For studies of organizational founding, the unit of analysis cannot be the organization because of the impossibility of defining the risk set. The set of social actors that are at risk of founding a hotel chain is undefined. A larger, definable, unit of analysis must be used. Founding rates increased because of the increases in the supply of professional managers and decreases in agency cost that chain entrepreneurs enacted. The founding models are further satisfying because all of the control variables were significant in the predicted directions. For the control variables, there is some disagreement between the implications of the founding and failure models. The non-monotonic effect of density is as predicted by the theory of density dependence. This is a little surprising because density had a monotonic effect on failure. The positive coefficient for Aggregate Franchise Units further supports the idea that franchising was important to the rise of hotel chains.
