This project uses economic methods to investigate how trading institutions affect prices and efficiency. We examine markets with directed search: buyers visit sellers based on any information they have. Simultaneous buyer choices and sellers' capacity constraints lead to 'frictions' where not all profitable exchanges occur - more realistic than the usual 'frictionless' assumption. We vary (a) whether sellers can post prices in advance; and (b) whether and how negotiation occurs based on how many buyers (1 vs. 2+) visit a seller. We use results from auction, bargaining, game and search theories, and new analysis, to form predictions, which we test using experiments. Our results should have implications for labour and competition policies.