We propose that we estimate geographic markets in two steps. First, estimate clusters of transactions interchangeable in use. Second, estimate markets from these clusters. We argue that these clusters are subsets of markets. We draw on both antitrust cases and economic intuition. We model and...